International Journal of Heat and Mass Transfer 93 (2016) 379387

Contents lists available at ScienceDirect

International Journal of Heat and Mass Transfer

journal homepage: www.elsevier.com/locate/ijhmt

An experimental study on heat transfer between supercritical carbon

dioxide and water near the pseudo-critical temperature in a double pipe
heat exchanger
Ting Ma a, Wen-xiao Chu a, Xiang-yang Xu a, Yi-tung Chen b, Qiu-wang Wang a,
a
b

Key Laboratory of Thermo-Fluid Science and Engineering, MOE, Xian Jiaotong University, Xian, Shaanxi 710049, China
Department of Mechanical Engineering, University of Nevada, Las Vegas, NV 89154-4027, USA

a r t i c l e

i n f o

Article history:
Received 16 April 2015
Received in revised form 6 October 2015
Accepted 9 October 2015
Available online 11 November 2015
Keywords:
Supercritical carbon dioxide
Heat transfer
Double pipe heat exchanger
Correlation
Buoyancy force

a b s t r a c t
Supercritical carbon dioxide (SCO2) is a promising working fluid for the cryogenic refrigeration, aircondition and heat pump systems. The present study sets up a SCO2water test loop to study the heat
transfer performance of SCO2 in a double pipe heat exchanger. The effects of SCO2-side pressure, mass
flux and buoyancy force as well as water-side mass flux are investigated. It is found that the total and
SCO2-side heat transfer coefficients reduce as the SCO2-side pressure increases. The peak total and
SCO2-side heat transfer coefficients appear at a higher temperature than the pseudo critical temperature.
The water-side mass flux has a larger effect on the total heat transfer coefficient compared to the
SCO2-side mass flux in the studied cases. The contribution of buoyancy force to the heat transfer
performance is large at the small SCO2-side mass flux and it becomes smaller as the SCO2-side mass flux
increases. The SCO2-side pressure and water-side mass flux have little effect on the buoyancy force. A
heat transfer correlation that includes the effect of buoyancy force is obtained by fitting the experimental
data with genetic algorithm.
2015 Elsevier Ltd. All rights reserved.

1. Introduction
The halohydrocarbon refrigerants such as CFCs and HCFCs were
restricted for use in many years ago due to the damage to ozone
layer [1]. The widely used HFC-134a is also not recommended to
be used in the future because it may increase the greenhouse
effect, and may be decomposed into acid and other toxic substance
by the sunniness in the troposphere [23]. Recently, many
researchers have paid considerable attention to the CO2. The CO2
is non-toxic, incombustible and safe so that it almost has nonnegative effects on the environment whose Ozone Depletion
Potential (ODP) and effective Global Warming Potential (GWP)
are zero [45]. When the pressure and temperature approach the
critical point of CO2, the specific heat capacity is very high which
can significantly improve the heat transfer performance. The
supercritical fluids cannot be defined as a liquid or as a gas but
as a substance in a state, i.e., supercritical state, because their thermal physical properties are different from those of real fluids, and
there is no liquidvapor phase transition and interfaces at supercritical pressures. Therefore, the trans-critical CO2 system has been
Corresponding author. Tel.: +86 29 82665539; fax: +86 29 82663502.
E-mail address: wangqw@mail.xjtu.edu.cn (Q.-w. Wang).
http://dx.doi.org/10.1016/j.ijheatmasstransfer.2015.10.017
0017-9310/ 2015 Elsevier Ltd. All rights reserved.

used for the cryogenic refrigeration, air-condition and heat pump.

It was found that the optimized trans-critical CO2 system could get
comparative performance compared to the synthetic refrigerants
system in the commercial light cryogenic refrigeration [67]. The
performance of heat pump using trans-critical CO2 system was better than that using traditional refrigerants because it could heat the
water up to 90 C even at the atmosphere being
20 C, and
reduced the CO2-side energy loss [8]. The trans-critical CO2 heat
pump system proposed by the energy department of Norway could
operate under three conditions, i.e., only room heating, only hot
water supply, and both room heating and hot water supply [9].
However, the operating pressure of trans-critical CO2 system is
very high so that there is a big challenge for the safety. As one of
the key components, the heat exchanger plays an important role
to transfer the heat from one loop to the other loop. The heat
exchanger for trans-critical CO2 system should endure the extreme
high pressure. Besides, the thermal-hydraulic performance of heat
exchanger has a significant effect on the efficiency, compactness
and cost of the system. Therefore, it is necessary to study the
thermal-hydraulic performance of CO2 in the high pressure heat
exchanger at supercritical pressures.
Considerable research studies on CO2 at supercritical pressures
have been conducted, especially after the SCO2 was proposed as

380

T. Ma et al. / International Journal of Heat and Mass Transfer 93 (2016) 379387

Nomenclature
A
c1, c2, c3
D
f
g
Gm
Gr
h
H
L
Nu
p
Q
qm
Re
Pr
T
U

heat transfer area, m2

constants in correlation
diameter, m
Darcy friction factor
gravitational acceleration, m s
2
mass flux, kg m
2 s
1
Grashof number
heat transfer coefficient, W m
2 K
1
enthalpy, J kg
1
tube length, m
Nusselt number
pressure, MPa
heat transfer rate, W
mass flow rate, kg s
1
Reynolds number
Prandtl number
temperature, C
total heat transfer coefficient, W m
2 K
1

one of working fluids in the nuclear reactor in 1950s. In 1967, Hall

et al. [10] compared the available experimental data with the
empirical formulas and found that more attention should be paid
on the buoyancy force even in the forced convection condition.
Pioro et al. [11] applied many empirical formulas to calculate the
heat transfer coefficient of SCO2 in the tube under the same operating condition. There were significant differences both in the values and trends among the different empirical formulas, which
indicated that the empirical formulas used for the fluids with constant thermal physical properties could not include the effect of
remarkable variation of thermal physical properties. The heat
transfer performances of SCO2 in vertical tubes with circular, triangular and square cross sections were tested and compared [12].
Song et al. [13] found that the heat transfer had similar characteristics regardless of heat transfer enhancement or deterioration if
the ratio of tube length to tube diameter and the ratio of heat flux
to mass flux were kept as constant. Kim et al. [14] performed an
experiment to examine the effect of flow directions including the
up flow and down flow on the heat transfer performance. The
result showed that if the heat flux of wall was middle and mass
flux was small, the peak point of wall temperature appeared in
the up flow, but the wall temperature increased monotonously
along the flow direction in the down flow. Both the flow and heat
transfer were affected by the buoyancy force and flow acceleration.
Zhang et al. [15] tested the mixed convective heat transfer of SCO2
inside a vertical helically coiled tube. The result showed that the
buoyancy force, centrifugal force and variations of thermal physical properties had significant effects on the temperature and heat
transfer coefficient distributions along the circumference edges.
Yang et al. [16] numerically studied the cooling process of SCO2
in the tube at constant wall temperature. It showed that the contribution of inclined angles on heat transfer decreased as the gravity
force magnitudes reduced. Bae [17] conducted an experiment to
analyze the influence of flow direction and flow channel shape,
and evaluate the existing SCO2 heat transfer correlations. Chen
et al. [18] discovered multiple peaks of heat transfer coefficients
along the horizontal flow, and the effects of inlet Reynolds number,
channel diameter and heat transfer to trans-critical CO2 flow inside
the mini-channels were examined. Chen et al. [19] examined the
near-critical CO2 thermosyphon in the trans-critical/SCO2 based
natural circulation loop.

Greek
DT
Dp

q
k

logarithmic mean temperature difference, K

pressure difference, Pa
density, kg m
3
thermal conductivity, W m
1 K
1
dynamic viscosity, Pa s

Subscripts
b
average value at cross section
CO2
supercritical CO2 side
H2O
water side
in
inlet
out
outlet
pc
pseudo critical point
w
wall
Wall-H outer diameter
Wall-C inner diameter

However, the above researches mainly focused on the heat

transfer process of SCO2 under the constant heat flux or constant
wall temperature conditions. In the high pressure heat exchanger
used for the trans-critical CO2 system, there involves a heat
exchange process between the water to SCO2, whose thermal
boundary is different from the constant heat flux and thus may
result in a different heat transfer performance. Therefore, the present study conducts an experiment to study the heat transfer performance of SCO2 in a double pipe heat exchanger using SCO2
water test loop.
2. Experimental system
The experimental system is composed of two separate closed
loops, i.e., CO2 loop and water loop, as shown in Fig. 1. In the CO2
loop, the subcritical CO2 at 35 MPa released from the tank is
cooled to liquid by the refrigerator, and then injected into the pipeline of system using the gas driving liquid hydraulic pump. The gas
driving liquid hydraulic pump can provide the maximum pressure
of 30.2 MPa for the CO2. Therefore, the CO2 can reach the supercritical state by increasing the pressure of system. The mass flow rates
of water and CO2 on both sides are controlled according to the
micropump, and measured by the flowmeter. The maximum volume flow rate supported by micropump is 20 L min
1. The flowmeter can measure the mass flow rate ranged from 80 kg h
1 to
1500 kg h
1 with 0.05% accuracy. The maximum measuring pressure for the gauge pressure transducers is 27.6 MPa, while that
for the differential pressure transducers is 62.2 kPa. The accuracy
of gauge pressure transducers is 0.065%, and that of differential
pressure transducers is 0.075%. The inlet temperature of test section on both sides is controlled by the heaters, and the T-type thermocouples are used to measure the temperatures. The maximum
measuring temperature for the T-type thermocouples is 200 C.
The T-type thermocouples were calibrated ahead of the experiment and the uncertainty of the measured temperature is within
0.1 C accuracy in the range of the present tested operating
conditions.
The test section is a double pipe heat exchanger, as shown in
Fig. 2. The water flows from the bottom to the top in the annular
space, while the SCO2 flows from the top to the bottom in the inner
tube. The inner tube is made of 316L stainless steel, but the outer

T. Ma et al. / International Journal of Heat and Mass Transfer 93 (2016) 379387

381

Fig. 1. Experimental system.

U CO2

Q ave
ACO2 DT

where, Qave is the average heat transfer rate, ACO2 is the heat transfer area of SCO2 side, and DT is the temperature difference between
the hot and cold sides.
During the experiment process, the experimental data is considered to be valid when the relative error of heat transfer rates
between the hot and cold sides is less than 10% in normal operating
condition and 15% in the pseudo-critical operating condition. The
average heat transfer rate Qave is calculated by:

Q ave Q CO2 Q H2O =2

where, QH2O is the heat transfer rate on the water side, and QCO2 is
the heat transfer rate on the SCO2 side, which are calculated by:


Q H2O Gm;H2O Ain;H2O HH2O;out
HH2O;in



Q CO2 Gm;CO2 Ain;CO2 HCO2;in
HCO2;out

Fig. 2. Photo of test section: left side-test section without thermal insulation
material; right side-test section with thermal insulation material.

tube is made of transparent plexiglass tube for the convenience of

installing thermocouples along the streamwise direction of annular space. The outer surfaces of test section and pipelines in the
system are covered by the thermal insulation material. The length
of double pipe heat exchanger is 1500 mm. The outer tube has an
outer diameter of 25 mm with a thickness of 2 mm, while the inner
tube has an outer diameter of 16 mm with a thickness of 2 mm. In
order to investigate the local heat transfer performance, ten pairs
of thermocouples are mounted in the center of annular cross sections along the streamwise direction of water side, as shown in
Fig. 3. The distance of adjacent cross section is 100 mm and there
are two thermocouples in every cross section.

3. Data reduction
In the double pipe heat exchanger, the total heat transfer
coefficient UCO2 based on the SCO2-side heat transfer area is
calculated by:

3
4

where, HH2O,in and HH2O,out are the enthalpy of water at the inlet and
outlet, while HCO2,in and HCO2,out are the enthalpy of SCO2 at the inlet
and outlet. Ain,H2O and Ain,CO2 are the inlet cross-sectional area on
the H2O side and SCO2 side, respectively. The thermal physical
properties of SCO2 dependent with temperature and pressure are
considered during the calculation, which are obtained from the
NIST website [20].
It was found that the temperature difference between the
inlet and outlet on SCO2 side is less than 3 C because the tube
length and the temperature difference between water and SCO2
sides are relatively small. Thus, the logarithmic mean temperature difference is used to calculate the DT in the countercurrent
flow:

DT




T CO2;out
T H2O;in
T CO2;in
T H2O;out


T

T H2O;in
ln T CO2;out
CO2;in
T H2O;out

According to the thermal resistance network in the double pipe

heat exchanger, the total heat transfer coefficient UCO2 can also be
calculated by:

1
1
ACO2 ln DWall-H =DWall-C
ACO2

U CO2 hCO2
2pkL
hH2 O AH2 O

Then the heat transfer coefficient hCO2 on SCO2 side is calculated

by:

382

T. Ma et al. / International Journal of Heat and Mass Transfer 93 (2016) 379387

Fig. 3. Sketch of test section and locations of mounted thermocouples (unit: mm).

where, DWall-H is the outer diameter of inner tube, DWall-C is the

inner diameter of inner tube, and k is the thermal conductivity of
316L stainless steel.
From Eq. (7), it can be found that the heat transfer coefficient
hH2O in the annular space of water side should be obtained at first
before the calculation of heat transfer coefficient hCO2 in the inner
tube. The following three steps are adopted to calculate hH2O.
At first, a heat transfer experiment in the double pipe heat
exchanger using waterwater test loop is conducted by keeping
the mass flow rate on annular side constant, but changing the mass
flow rate on inner tube side. Then the Nusselt number of water on
the inner tube side can be obtained by comparing the experimental
data with Wilson Plots. The obtained Nusselt number and corresponding friction factor of water on the inner tube side are shown
in Fig. 4. Compared to the Filonenko [21] and Gnielinski [22] formulas for inner tube, the maximum relative deviation in Nusselt
number is less than 8%, and that in Darcy friction factor is less than
7%. The Gnielinski and Filonenko formulas are modified based on
the present experimental data for the later analysis, as shown in
follows:

Nu

 2=3
f =8Re
1000Pr
D
 1
p 2=3
L
1:17 12:31 f =8 Pr
1

f 1:078 lgRe 1:446

2

Present results
Gnielinski formula

180
160
140
120
100
80

Deviationmax = 7.2%

60
1.0x10

1.5x10

2.0x10

2.5x10

3.0x10

Re

(a) Comparison of Nusselt number on inner tube side

0.040

Present results
Filonenko formula

0.035

8
0.030

The maximum relative deviation in the modified Gnielinski

shown in Eq. (8) is less than 2%, and that in the modified Filonenko
formulas shown in Eq. (9) is less than 1.1% compared to the experimental data. It should be noted that Eqs. (8) and (9) are obtained
in the water-water test loop.
Secondly, a heat transfer experiment in the double pipe heat
exchanger using waterwater test loop is conducted by keeping
the mass flow rate on inner tube side at constant, but changing
the mass flow rate on annular side. Then the Nusselt number on
the annular side can be obtained based on the above obtained
Nusselt number on the inner tube side. As shown in Fig. 5, the
maximum relative deviation in Nusselt number is less than 6%,
and that in Darcy friction factor is less than 17% compared to the
empirical formulas for inner tube. Because there are 20 thermocouples along the streamwise direction of annular side, these thermocouples can disturb the flow and thus increase the friction
factor. Therefore, the measured Darcy friction factor is larger than
the empirical formula. However, the Darcy friction factor on the

Deviationmax = 6.14%

"

200

Nu



1
ACO2 ln DWall-H =DWall-C
ACO2
hCO2 1




U CO2
2pkL
hH2 O AH2 O

0.025

0.020

0.015
4
1.0x10

1.5x10

2.0x10

2.5x10

3.0x10

Re

(b) Comparison of friction factor on inner tube side

Fig. 4. Comparison of experimental result and empirical formulas on inner tube
side in waterwater test loop.

annular side is unnecessary to calculate the heat transfer coefficient on both sides. Therefore, the present experimental data is
valid. The fitting formulas based on the present experimental data
are shown as follows:

383

T. Ma et al. / International Journal of Heat and Mass Transfer 93 (2016) 379387

can be calculated by Eq. (7). Here, the heat transfer coefficient on

water side is calculated based on the Eq. (10). According to the
method proposed by Kline and McClintock [23], the experimental
uncertainty of the main variables is summarized in Table 1.

60
Present results
Gnielinski formula

55
50

4. Results and discussion

Nu

45
Deviationmax = 5.75%

4.1. Effect of SCO2-side pressure

40
35
30

5000

5500

6000

6500

Re

(a) Comparison of Nusselt number on annular side

0.08
Present results
Filonenko formula

0.07
0.06

Deviationmax = 17%

0.05
0.04
0.03
0.02
0.01
3500

4000

4500

5000

5500

6000

6500

Re

(b) Comparison of friction factor on annular side

Fig. 5. Comparison of experimental result and empirical formulas on annular side
in waterwater test loop.

Nu 0:012Re0:92 Pr0:21
f 1:23 lgRe 0:2

10

2

4.2. Effect of SCO2-side mass flux

The total heat transfer coefficient under different SCO2-side
mass flux conditions is shown in Figs. 8 and 9. Increasing the
SCO2-side mass flux can enhance the total heat transfer coefficient.
Specifically, the increase is more obvious at the region lower than
pseudo critical temperature when the SCO2-side mass flux is small,
but more remarkable at the region greater than pseudo critical
temperature when the SCO2-side mass flux is large. It can also be

11

The maximum relative deviation of Nusselt number predicted

by the fitting formula shown in Eq. (10) is less than 1.7%, and that
of friction factor predicted by the fitting formula shown in Eq. (11)
is less than 0.65% compared to the experimental data.
At last, the heat transfer experiment in the double pipe heat
exchanger is conducted in the SCO2water test loop. The effects
of inlet pressure and temperature on SCO2 side, mass flow rate
on both sides are studied. The total heat transfer coefficient is
obtained by Eq. (1). Then the heat transfer coefficient on SCO2 side

Table 1
Experimental uncertainty analysis.

2800
Tpc = 34.7C

Uncertainty (%)

Variables

Uncertainty (%)

p
T
qm
UCO2
Re

0.12
0.5
0.16
4.61
2.1

Dp
h
f
Nu
Q

0.72
6.52
6.17
6.54
4.51

pSCO2 = 9 MPa
pSCO2 = 10 MPa

2000

1600
Tpc = 45C
-2

-1

-2

-1

Gm, SCO2 = 491 kgm s

1200

Variables

pSCO2 = 8 MPa

Tpc = 40C

2400

-1

4500

-2

4000

UCO2 /Wm K

25
3500

The variations of thermal physical properties of SCO2 become

stronger when the operating pressure is close to the critical pressure of CO2. Fig. 6 shows the effect of SCO2-side pressure on the
total heat transfer coefficient. It can be seen that the variation
trend of total heat transfer coefficient is similar to that of specific
heat under different operating pressures. All the total heat transfer
coefficients reach to the peak values near the pseudo critical point
of SCO2. This is caused by the fact that the specific heat of SCO2 is
the maximum at the pseudo critical point so that the heat transfer
performance is the best. However, the peak value of total heat
transfer coefficient appears at the location whose temperature is
a little higher than the pseudo critical temperature. The reason
may be that the temperature of SCO2 near the wall is lower than
that in the center during the cooling process. Although the average
temperature of SCO2 in the cross section is not reduced to the
pseudo critical temperature, the temperature of SCO2 near the wall
has closely reached to the pseudo critical temperature. Therefore,
the temperature corresponding to the peak point of total heat
transfer coefficient shifts behind the pseudo critical temperature.
The effect of SCO2-side pressure on the SCO2-side heat transfer
coefficient is shown in Fig. 7. The SCO2-side heat transfer coefficient reduces obviously as the pressure increases, but the variation
trend becomes smoother. The SCO2-side heat transfer coefficient is
much higher than the total heat transfer coefficient under the same
operating condition.

Gm, H2O = 903 kgm s

800
20

25

30

35

40

45

50

55

60

65

Tb /C
Fig. 6. Effect of SCO2-side pressure on total heat transfer coefficient.

70

384

T. Ma et al. / International Journal of Heat and Mass Transfer 93 (2016) 379387

4.4. Effect of SCO2-side buoyancy force

9000
pSCO = 8 MPa

Tpc = 34.7 C

8000

pSCO = 9 MPa
2

Tpc = 40 C

7000

pSCO = 10 MPa
2

-2

hCO2 /Wm K

-1

6000

-2

-1

Gm, SCO 2 = 707 kg m s

5000

-2

Gm, H2O = 1670 kg m s

-1

4000
3000
Tpc = 45 C

2000
1000
20

25

30

35

40

45

50

55

60

65

70

75

Tb /C

It was found by Pioro et al. [11] that the buoyancy force has a
significant effect on the heat transfer performance of SCO2 in the
circular tube during the heating process. According to the different
effects of buoyancy force on the flow and heat transfer performance, the flow of SCO2 can be divided into forced convection
and mixed convection. In the forced convection process, the effect
of buoyancy force is very small, but that of thermal physical properties of SCO2 is significant. In the mixed convection process, both
the buoyancy force and thermal physical properties of SCO2 play
major roles on the heat transfer performance, which may result
in the heat transfer deterioration or heat transfer enhancement.
Jackson and Hall [24] pointed out that the buoyancy force of
SCO2 played an important role and could not be neglected when
the buoyancy force satisfied the following equation.

Gr
Re2:7

Fig. 7. Effect of SCO2-side pressure on SCO2-side heat transfer coefficient.

> 10
5

12

where Gr is the Grashof number, which is defined as follows:

Gr

3200
-2

-1

-2

-1

-2

-1

Gm, SCO2 = 491 kgm s

Gm, SCO2 = 627 kgm s

-2

UCO2 /Wm K

-1

2800

 W qb gD3
qb
q

One major reason for buoyancy force to affect the heat transfer
performance is that the large radial temperature gradient makes a
large density gradient. In Eq. (13), the density gradient is characterized by the difference of the density qb at the fluid average temper w is
 w at the wall average temperature. q
ature and the density q
calculated by:

Gm, SCO2 = 823 kgm s

2400

2000

qw
-2

-1

Gm, H2O = 903 kgm s

1200
20

25

30

35

40

qw qb =2
8 T w > T pc ; or; T b < T pc
qb T b
T pc qw T pc
T w
8 T w < T pc < T b
T
T w

14

pSCO2 = 8 MPa

1600

13

l2b

45

50

55

Tb /C
Fig. 8. Effect of SCO2-side mass flux on total heat transfer coefficient.

found that the contribution of SCO2-side heat transfer coefficient to

the total heat transfer coefficient is large when the SCO2-side mass
flux is small, but becomes smaller as the SCO2-side mass flux keeps
increasing.

Fig. 12 shows the comparison of SCO2-side buoyancy force

along the streamwise direction under different pressure conditions. The buoyancy force increases at first and then reduces,
whose peak value locates at the pseudo critical temperature. The
difference of buoyancy force under different operating pressures
only appears when the temperature is greater than the pseudo critical temperature but becomes small when the temperature keeps
increasing. The pressure also has little influence on the buoyancy
force when the temperature is lower than the pseudo critical temperature. According to Eq. (12), the contribution of buoyancy force
on the heat transfer performance under the studied operating

16000

4.3. Effect of water-side mass flux

14000

-2 -1
-2 -1

Gm, SCO2 = 823 kgm s

-2 -1

Gm, H2O = 903 kgm s

-1
-2

Gm, SCO2 = 627 kgm s

pSCO2 = 8 MPa

12000
hCO2 /W m K

The effects of water-side mass flux on the total and SCO2-side

heat transfer coefficients are shown in Figs. 10 and 11. The total
and SCO2-side heat transfer coefficients increase with the increase
of water-side mass flux, whose trend is similar to the parabola
shape. However, the increase in the total heat transfer coefficient
is much larger compared to that in the SCO2-side heat transfer
coefficient. Increasing the water-side mass flux increases the
water-side heat transfer coefficient so that the total heat transfer
coefficient increases, but has a minor effect on the wall temperature of inner tube so that SCO2-side heat transfer coefficient
changes little. Compared to Fig. 8, it can be found that the waterside mass flux has a larger effect on the total heat transfer coefficient than the SCO2-side mass flux. Therefore, the heat transfer
performance of double pipe heat exchanger is more determined
by the water-side heat transfer performance than the SCO2-side
heat transfer performance in the SCO2water test loop in the studied operating conditions.

-2 -1

Gm, SCO2 = 491 kgm s

10000
8000
6000
4000
2000
0
20

25

30

35

40

45

50

Tb /C
Fig. 9. Effect of SCO2-side mass flux on SCO2-side heat transfer coefficient.

55

385

T. Ma et al. / International Journal of Heat and Mass Transfer 93 (2016) 379387

1x10-3

3200
-2

-1

-2

-1

Gm, H2O = 536 kgm s

pSCO = 8 MPa
2

Gm, H2O = 903 kgm s

2800

-2

pSCO = 10 MPa
2

2400
Gr/Re

-2

2.7

-1

1x10-4

-1

Gm, H2O = 1128 kgm s

UCO2 /Wm K

pSCO = 9 MPa

2000

1x10-5
Gm, SCO2 = 491 kgm-2s-1

1600

1x10

-6

Gm, H2O = 903 kgm-2s-1

pSCO2 = 8 MPa

1200

-2

-1

Gm, SCO2 = 491 kgm s

1x10-7
0.6

800
20

25

30

35

40

45

50

55

60

0.8

1.0

1.2

1.4

Tb (C)

1x10-3

1x10

Bruch et al. [25] found that the modified Krasnoshchekov

correlation proposed by Jackson and Hall [26] could give a good
prediction for the forced convection process when the buoyancy
force was ignored in the cooling process of interior flow. The
modified Krasnoshchekov correlation is shown as follows:

Gm, SCO2 = 491 kgm-2s-1

Gm, SCO2 = 823 kgm-2s-1

-4

1x10-5

pSCO2 = 8 MPa

1x10-6

4.5. SCO2-side heat transfer correlation

Gm, H2O = 903 kgm-2s-1

1x10-7
0.6

0.8

1.0

1.2

-3

Gm, H2O = 536 kgm-2s-1

-2 -1

9000
8000

Gm, H2O = 903 kgm s

1.6

Fig. 13. Comparison of SCO2-side buoyancy force along streamwise direction under
different SCO2-side mass flux conditions.

1x10
Gm, H2O = 536 kgm s

1.4

Tb/Tpc

10000

Gm, H2O = 903 kgm-2s-1

-2 -1

Gm, H2O = 1128 kgm-2s-1

-2 -1

Gm, H2O = 1128 kgm s

-4

Gr/Re2.7

1x10

-1
-2

2.0

Gm, SCO2 = 627 kgm-2s-1

Gr/Re2.7

pressure conditions is remarkable when the mass flux is

491 kg m
2 s
1 and the average temperature is smaller than
1.5Tpc, which belongs to the mixed convection.
As shown in Fig. 13, the SCO2-side mass flux significantly affects
the SCO2-side buoyancy force and thus changes the heat transfer
performance. The contribution of SCO2-side buoyancy force to heat
transfer performance reduces as the SCO2-side mass flux increases,
which can be neglected when the SCO2-side mass flux is greater
than 823 kg m
2 s
1 because the flow is in the forced convection
range. However, the water-side mass flux has little influence on
the SCO2-side buoyancy force, as shown in Fig. 14. The contribution
of buoyancy force on the heat transfer performance is remarkable
when the average temperature is smaller than 1.4Tpc  1.5Tpc
under the studied mass flux conditions.

hCO2 /Wm K

1.8

Fig. 12. Comparison of SCO2-side buoyancy force along streamwise direction under
different pressure conditions.

Fig. 10. Effect of water-side mass flux on total heat transfer coefficient.

7000

1.6

Tb/Tpc

6000
5000

1x10-5

4000

pSCO2 = 8 MPa

3000

Gm, SCO2 = 491 kgm-2s-1

2000

pSCO2 = 8 MPa

1000

Gm, SCO2 = 491 kgm s

-2 -1

1x10

-6

0.6

0
20

25

30

35

40

45

50

55

60

65

Tb (C)
Fig. 11. Effect of water-side mass flux on SCO2-side heat transfer coefficient.

0.8

1.0

1.2

1.4

1.6

1.8

Tb/Tpc
Fig. 14. Comparison of SCO2-side buoyancy force along streamwise direction under
different water-side mass flux conditions.

386

T. Ma et al. / International Journal of Heat and Mass Transfer 93 (2016) 379387

1600
+30%

1400

Nucorrelated

1200
1000
800

-30%

600
400
200
200

400

600

800

1000

1200

1400

1600

Nuexperimental

(2) The total and SCO2-side heat transfer coefficients increase as

the SCO2-side mass flux increases. But the changes become
smaller as the SCO2-side mass flux keeps increasing. The
water-side mass flux has a larger effect on the heat transfer
performance compared to SCO2-side mass flux in the studied
cases.
(3) The influence of buoyancy force on the heat transfer performance is remarkable near the pseudo critical temperature
because the physical properties of SCO2 change significantly
in this temperature region. The contribution of buoyancy
force to the heat transfer performance is large at the small
SCO2-side mass flux and becomes smaller as the SCO2-side
mass flux increases. The SCO2-side pressure and water-side
mass flux have little effect on the buoyancy force. The
experimental correlation is fitted by genetic algorithm,
which includes 90% of experimental data within 30%
error bars.
Conflict of interest

Fig. 15. Comparison of experimental data and fitting correlation.

None declared.


0:5
NuFc 0:0183Re0:82
b Pr b

qb
qW


0:3
15

Based on the above correlation and his experimental data,

Bruch et al. [25] proposed the following correlation to consider
the effect of buoyancy force:

8
 0:46
>
< 1
75 Gr2:7
Nub
Re

 0:4
NuFc >
: 13:5 Gr
2:7

Re

Gr
Re2:7

< 4:2  10
5

Gr
Re2:7

> 4:2  10
5

16

According to Figs. 1214, the Gr/Re2.7 is mainly between 10
5

and 4  10
5 because of the large tube diameter and SCO2-side
mass flux. Therefore, the following equation form is used:


c3
Nub
Gr
c1 c2
NuFc
Re2:7

17

The genetic algorithm is adopted to fit the correlation, as shown

in Eq. (18). The comparison of experimental data and fitting correlation is shown in Fig. 15. It can be seen that 90% of experimental
data are within 30% error bars.


0:458
Nub
Gr
2:61
86:965
NuFc
Re2:7

18

where 8:0  104 < Re < 4:9  105 and 11 < Pr < 130.
5. Conclusions
In this paper, an experiment is performed to study the heat
transfer performance of SCO2 in the double pipe heat exchanger
using the SCO2water test loop. The effects of SCO2-side pressure,
SCO2-side mass flux, water-side mass flux and SCO2-side buoyancy
force are studied. The main conclusions are summarized as
follows:
(1) The variation of total and SCO2-side heat transfer coefficients with temperature behaves like a parabola shape,
which is similar to the specific heat. However, the peak
points of total and SCO2-side heat transfer coefficients
appear at the temperature which is a little higher than the
pseudo critical temperature due to the large temperature
gradient in the radial direction. The total and SCO2-side heat
transfer coefficients reduce with the increase of SCO2-side
pressure.

Acknowledgements
This work is supported by the International Cooperation and
Exchanges Project of NSFC of China (Grant No. 51120165002)
and the National Natural Science Foundation of China (Grant No.
51306139), the Special Financial Grants from the China Postdoctoral Science Foundation (No. 2014T70919) and the Fundamental
Research Funds for the Central Universities of China.
References
[1] M. Protocol, Montreal Protocol on Substances that Deplete the Ozone Layer, US
Government Printing Office, Washington, DC, 1987. pp. 26.
[2] R.E. Banks, Scepticism about R134a justified, Refrig. Air Condit. 96 (1993) 16.
[3] C. Breidenich, D. Magraw, A. Rowley, J.W. Rubin, The Kyoto protocol to the
United Nations framework convention on climate change, Am. J. Int. Law 92
(1998) 315331.
[4] S.B. Riffat, C.F. Afonso, A.C. Oliveira, D.A. Reay, Natural refrigerants for
refrigeration and air-conditioning systems, Appl. Therm. Eng. 17 (1) (1997)
3342.
[5] G. Lorentzen, The use of natural refrigerants: a complete solution to the CFC/
HCFC predicament, Int. J. Refrig. 18 (1995) 190197.
[6] A. Pearson, Carbon dioxide-new uses for an old refrigerant, Int. J. Refrig. 28
(2005) 11401148.
[7] G. Boccardi, N. Calabrese, G.P. Celata, R. Mastrullo, A.W. Mauro, A. Perrone, R.
Trinchieri, Experimental performance evaluation for a carbon dioxide light
commercial cooling application under transcritical and subcritical conditions,
Appl. Therm. Eng. 54 (2) (2013) 528535.
[8] M. Saikawa, K. Hashimoto, T. Kobayakawa, K. Kusakari, M. Itoh, H. Sakakibara,
Development of prototype of CO2 heat pump water heater for residential use,
Refrig. Sci. Technol. (2001) 97102.
[9] J. Stene, Residential CO2 heat pump system for combined space heating and
hot water heating, Int. J. Refrig. 28 (2005) 12591265.
[10] W. Hall, J. Jackson, A. Watson, A review of forced convection heat transfer to
fluids at supercritical pressures, Proc. Inst. Mech. Eng. 39 (1967) 1022.
[11] I.L. Pioro, R.B. Duffey, Heat transfer and hydraulic resistance at supercritical
pressures in power engineering applications, American Society of Mechanical,
New York, 2007.
[12] J.K. Kim, H.K. Jeon, J.S. Lee, Wall temperature measurement and heat transfer
correlation of turbulent supercritical carbon dioxide flow in vertical circular/
non-circular tubes, Nucl. Eng. Des. 237 (2007) 17951802.
[13] J.H. Song, H.Y. Kim, H. Kim, Y.Y. Bae, Heat transfer characteristics of
a supercritical fluid flow in a vertical pipe, J. Supercrit. Fluid 44 (2008)
164171.
[14] D.E. Kim, M.H. Kim, Experimental investigation of heat transfer in vertical
upward and downward supercritical CO2 flow in a circular tube, Int. J. Heat
Fluid Flow 32 (2011) 176191.
[15] W. Zhang, S.X. Wang, C.D. Li, J.L. Xu, Mixed convective heat transfer of CO2 at
supercritical pressures flowing upward through a vertical helically coiled tube,
Appl. Therm. Eng. 88 (2015) 6170.
[16] C.Y. Yang, J.L. Xu, X.D. Wang, W. Zhang, Mixed convective flow and heat
transfer of supercritical CO2 in circular tubes at various inclination angles, Int.
J. Heat Mass Transfer 64 (2013) 212223.

T. Ma et al. / International Journal of Heat and Mass Transfer 93 (2016) 379387

[17] Y.Y. Bae, Mixed convection heat transfer to carbon dioxide flowing upward and
downward in a vertical tube and an annular channel, Nucl. Eng. Des. 241
(2011) 31643177.
[18] L. Chen, Y.M. Chen, M.H. Sun, X.R. Zhang, Investigation of trans-critical CO2
horizontal mini-channel flow with multi-peak heat transfer behaviors, Ann.
Nucl. Energy 75 (2015) 559569.
[19] L. Chen, B.L. Deng, X.R. Zhang, Experimental investigation of CO2
thermosyphon flow and heat transfer in the supercritical region, Int. J. Heat
Mass Transfer 64 (2013) 202211.
[20] NIST web-page, <http://webbook.nist.gov/chemistry/fluid>.
[21] G.K. Filonenko, On friction factor for a smooth tube, All Union Thermotechnical
Institute (Izvestifa VTI, No 10), Russia, 1948.
[22] V. Gnielinski, New equations for heat and mass-transfer in turbulent pipe and
channel flow, Int. Chem. Eng. 16 (1976) 359368.

387

[23] S.J. Kline, F. McClintock, Describing uncertainties in single-sample

experiments, Mech. Eng. 75 (1953) 38.
[24] J.D. Jackson, W.B. Hall, Influences of buoyancy on heat transfer to fluids
flowing in vertical tubes under turbulent conditions, in: S. Kakac, D.B. Spalding
(Eds.), Turbulent Forced Convection in Channels and Bundles, Hemisphere,
New York, 1979, pp. 613640.
[25] A. Bruch, A. Bontemps, S. Colasson, Experimental investigation of heat transfer
of supercritical carbon dioxide flowing in a cooled vertical tube, Int. J. Heat
Mass Transfer 52 (2009) 25892598.
[26] J.D. Jackson, W.B. Hall, Forced convection heat transfer to fluids at supercritical
pressure, in: S. Kakac, D.B. Spalding (Eds.), Turbulent Forced Convection in
Channels and Bundles, Hemisphere, New York, 1979, pp. 563611.