International Journal of Heat and Mass Transfer 92 (2016) 665674

Contents lists available at ScienceDirect

International Journal of Heat and Mass Transfer

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

Research on convection heat transfer character of super critical carbon

dioxide ows inside horizontal tube
Tao Ding, Zhen Li
Department of Engineering Mechanics, Tsinghua University, Meng Min Wei Technology Building, Beijing 100084, China

a r t i c l e

i n f o

Article history:
Received 3 May 2015
Received in revised form 8 September 2015
Accepted 9 September 2015
Available online 25 September 2015
Keywords:
Carbon dioxide
Pseudo critical point
Super critical
Heat transfer coefcient
Method of thermal resistance

a b s t r a c t
This paper mainly studied the heat transfer character of super critical carbon dioxide, also named R744.
Super critical carbon dioxide can be used in gas cooler for air conditioning system. Compared with traditional Freon coolant, R744 has higher density and lower viscosity, which is benet for heat transfer process. So it is important to study heat transfer character for super critical carbon dioxide. The experiment
research is taken for the temperature ranging from 29 to 55 C, while pressure is 8, 9 and 10 MPa. The
Reynold number is about 2  105. Thermal resistance method is used to measure heat transfer coefcient.
The results found that convection heat transfer enhances near the pseudo critical point, in the region far
away from pseudo-critical point, heat transfer character is just like ordinary single-phase uid. The heat
transfer coefcient violently changeable region is also the region where thermal property change rapidly.
It is also found that heat transfer coefcient gets its max at the region near the pseudo-critical point. In
addition, the results obtained in this paper are compared with other researchers results.
2015 Elsevier Ltd. All rights reserved.

1. Introduction
Ozone layer depletion and greenhouse effect are the two serious
environmental problems, which are caused mainly by Freon coolant like R22. So forbidding the use of Freon and searching for environmental friendly coolant seem to be serious problems. As a
result, reducing the carbon footprint and protecting environment
come more and more popular in the world [1].
Carbon dioxide, is an environmental friendly gas existing in the
atmosphere. Its ODP (ozone depression potential) equals to 0, and
GWP (global warming potential) equals to 1. It can be used widely
in industry, such as working uid in Rankine cycle [2], refrigeration
cycle [34] and seperated heat pipe system [57]. In the heat
exchanger of Rankine or refrigeration cycle, carbon dioxide is super
critical uid, so it is important to study the heat transfer character
of super critical carbon dioxide.
Researchers from all over the world have done a lot of work on
super critical carbon dioxide heat transfer character. Yoon [8] and
Son [9] study heat transfer character and pressure drop of super
critical carbon dioxide ows in tube in tube counter ow heat
exchanger, where carbon dioxide is cooled by water. Liao [10,11]
study both heating and cooling condition when super critical carbon dioxide ows in mini tube, its tube diameter is 0.7 mm,
Corresponding author. Tel.: +86 10 62772918.
E-mail address: lizh@tsinghua.edu.cn (Z. Li).
http://dx.doi.org/10.1016/j.ijheatmasstransfer.2015.09.026
0017-9310/ 2015 Elsevier Ltd. All rights reserved.

1.4 mm, 2.16 mm. Huai [12] does an experiment on ow and heat
transfer of super critical carbon dioxide in multi-port mini channels under cooling condition. Jiang [13,14] has done some work
on convection heat transfer when carbon dioxide ows in vertical
mini-tubes and in porous media. Table 1 shows some recent work
on super critical carbon dioxide convection heat transfer in horizontal tube.
Traditional method to test the heat transfer coefcient is to use
the method of denition, which means h = q/(Tf  Twall). In this
equation, q is the heat ux, Twall means the inside wall temperature
and Tf means the uids bulk mean temperature. Twall is calculated
by one dimension heat transfer equation after getting the outside
wall temperature. Tf is tested by thermal couples which is inserted
into the tube. This is a simple method for its calculation equation is
simpler. Literature [13] use the similar method, there is a little
difference, the uid bulk mean temperature is not tested but calculated by thermal equilibrium. Literature [17] also use this method,
the difference is that temperature difference in not Tf  Twall, but
the LMTD, where it is dened by R744s inlet/outlet temperature
and the inside wall temperature of each little section of the test
section.
However, there are some shortcomings for this method. First, as
for the super critical R744, its pressure is as high as 810 MPa,
inserting thermal couples into the tube can be dangerous, besides,
it is also difcult to seal. Second, the inside tube diameter of our
experiment is 7.5 mm, which is very small, inserting thermal

666

T. Ding, Z. Li / International Journal of Heat and Mass Transfer 92 (2016) 665674

Nomenclature
Cp
h
LMTD
Nu
Pc
Pr
Q
q
R
Re
T

specic heat, J/kg C

heat transfer coefcient, W/m2 K
log mean temperature difference
Nusselt number
pseudo critical point
Prandtl number
heat capacity, W
heat ux, W/m2
thermal resistance, K/W
Reynold number
temperature, C

Greek symbols
q
density, kg/m3
k
heat conductivity, W/m K
l
viscosity, kg/ms
Subscripts
i
enthalpy
m
mean
max
maximal
min
minimum

couples into the tube may inuence the ow eld. Third, it may not
be accurate to use one dimension heat transfer equation to calculate the inside wall temperature after getting the outside wall
temperature.
Besides, there is also another method to nd the heat transfer
coefcient [11]. In this method, there is a tube in tube counter ow
heat exchanger where hot R744 ows inside the tube and cold
water ows outside the tube. There are 6 thermal couples to measure the outside wall temperature of the inside tube (R744 tube).
The author try to keep the out wall temperature of the inside tube
as uniform as possible. The author uses the average of the 6 thermal couples as the outside wall temperature, and calculate the
inside wall temperature. Finally, they use h = Q/A.LMTD. Where A
is the heat transfer area, and LMTD is the log mean temperature
difference dened by the R744s inlet/outlet temperature and the
inside tube wall temperature. However, this method also has a little problem, for it use the average of the 6 thermal couples as the
outside wall temperature. In fact, cold water is used to cool the hot
R744, as a result, there exists temperature difference of these 6
thermal couples, thus, it may be not very accurate to regard the
outside wall temperature as constant.
In this paper, a different method to measure heat transfer coefcient is used, we called it method of thermal resistance. Literature
[18] has used a similar method to determine in-tube evaporation
heat transfer of carbon dioxide, in this paper, we have modied
thermal resistance method which is different from literature
[18]. This method has some advantages, for example, it is easy to

do the experiment, for thermal couples only need to be put at

the inlet/outlet of cold water and carbon dioxide. Thermal couples
do not need to be inserted into the inside tube, thus, the problems
of seal and dangerous can be solved. Besides, heat transfer coefcient is just calculated by the experiment results directly, there is
nearly no assumption on calculating the results. As a result, this
method is accurate, simple and safe. Super critical carbon dioxide
ows in a tube in tube counter ow heat exchanger, where it is
cooled by water. The diameter of the tube is 7.5 mm. Our research
is taken for the temperature ranging from 29 to 55 C, while pressure is 8, 9 and 10 MPa. The Reynold number is about 2  105.

2. Theoretical analysis of super critical carbon dioxide heat

transfer
2.1. Physical property of super critical carbon dioxide
If temperature and pressure are lower than critical point, the
uid is ordinary. At a certain pressure, when absorbing heat, the
uid change its phase from liquid to two-phase and nally to
gas. If temperature and pressure are higher than critical point,
we call such uid super critical uid.
There are mainly 2 signicant characters for the super critical
uid.
First is that there is no phase change for super critical uid.
Super critical phase is a kind of uniform single phase. Its viscosity

Table 1
Researches on super critical carbon dioxide convection heat transfer.
Researcher

Heat exchanger type

Temperature range (C)

Pressure range (MPa)

Inside tube diameter (mm)

Yoon [8]
Son [9]
Liao [10]
Liao [11]
Huai [12]
Pettersen [15]
Dang [16]
Lv [17]

Tube in tube Water cold

Tube in tube Water cold
Electric heating pipe
Tube in tube Water cold
Multi-port mini channels
Multi-port mini channels
Tube in tube Water cold
Tube in tube Water cold

30  65
20  90
20  110
20  110
22  53
15  70
30  70
30  65

7.5  8.8
7.5  10
7.4  12
7.4  12
7.4  8.5
8  10.1
8  10
8  10

7.73
7.75
0.7, 1.4, 2.16
0.7, 1.4, 2.16
1.31
0.79
1, 2, 4, 6
3.8

Table 2
Thermal properties comparable table of R744.
Fluid type(R744)

Condition

Density (kg/m3)

Viscosity (105 kg=ms)

Heat difussion rate (108 m2 =s)

Liquid
Gas
Super critical

5 MPa, x = 0
5 MPa, x = 1
10 MPa, 40 C

827.3
156.7
630

7.871
1.674
4.774

3.242
5.825
1.983

667

T. Ding, Z. Li / International Journal of Heat and Mass Transfer 92 (2016) 665674

is as sparse as gas, while its density is as heavy as liquid, as can be

seen in Table 2.
Second, thermal properties change quiet violently with temperature and pressure, especially temperature. As can be seen in Fig. 1.
x-axis is temperature and y-axis means dimensionless thermal
properties. For example, in Fig. 1(a), the value of density at 30 C
is calculated by q30  C =qpc . When pressure is 8 MPa, take the change
tendency of specic heat for example, specic heat at 35 C
(pseudo-critical point is 34.7 C) is nearly 9 times bigger than
specic heat at 20 C. The variation trends of thermal property
comes gentler with the increase of pressure. At 15 MPa, specic
heat at 64 C (pseudo-critical point) is only 30% times bigger than
specic heat at 40 C.
Now, we can do some theoretical analysis about the heat transfer coefcient. According to convection heat transfer theory [19],
heat transfer coefcient is the function of six parameters, which
are thermal conductivity, viscosity, density, specic heat, ow rate

and inside tube diameter. Thus, h = f (k,l,q,cp,t,d). Thermal conductivity, viscosity, density, specic heat are thermal properties, which
can be only determined by temperature and pressure. If we
identify the Re number and Pr number, then we can use Eq. (2)
to determine the heat transfer coefcient. P is pressure, and T is
temperature.

Re

ul

Pr

tqcp
k

t
hl
; Nu
k
a

Nu f Re; Pr P; T; u; l

We take the diameter of the inside tube as characteristic length

(l = d), If the tube diameter (d) and ow rate (u) are constant, then
the heat transfer coefcient is only the function of thermal properties. When temperature and pressure change, thermal properties
are also changed, which leads to the change of heat transfer coefcient. In the region where thermal properties change slowly (like

density

2.5

conductivity

Cp

viscosity

2
1.5
1
0.5
0
20

25

30

35

40

45

50

55

60

65

70

T/oC

(a) P=8MPa
3

density

2.5

conductivity

Cp

viscosity

2
1.5
1
0.5
0
20

25

30

35

40

45
T/oC

50

55

60

65

70

(b) P=10MPa
3

density

2.5

conductivity

Cp

viscosity

2
1.5
1
0.5
0
40

45

50

55

60

65
T/oC

70

75

80

85

(c) P=15MPa
Fig. 1. The variation trend of thermal physical properties with temperature and pressure (R744).

90

668

T. Ding, Z. Li / International Journal of Heat and Mass Transfer 92 (2016) 665674

ordinary single phase uid), the heat transfer coefcient also

changes slowly. However, in the region thermal properties change
violently (like super critical carbon dioxide), the heat transfer coefcient also changes quite a lot.
3. Introduction of thermal resistance method
In a counter ow heat exchanger, hot carbon dioxide ows in
the inside tube and cold water ows in the outside tube.
Heat needs to transfer from the inside hot carbon dioxide to the
outside cold water. If we identify the temperature difference as the
Logarithmic mean heat transfer temperature difference (LMTD),
see Eq. (6), then the total heat exchange capacity can be calculated
by Newton cooling law (Eq. (3)). R means total thermal resistance,
1/R can be seen as the total heat transfer coefcient. Apparently, at
a certain temperature difference, the less R results to stronger heat
exchange capacity. There are 3 thermal resistance need to be considered when heat transfers from hot carbon dioxide to cold water.
They are inside tube convection thermal resistance R1, conduction
through tube wall thermal resistance R2 and outside convection
thermal resistance R3. R is the sum of R1, R2, R3.

Dt m
R

Fig. 3 shows the scheme of the test section. It is a tube in tube

counter ow heat exchanger, hot carbon dioxide ows in the inside
tube, and it is cold by water outside the tube. There is heat insulated
material outside the test section, so we can believe that heat is all
transferred from hot carbon dioxide to the cold water. In other
word, no heat loss from the test section. Besides, the tube is surrounded by heat insulated material, so there is no heat ux through
the wall of the inside tube, as is shown in Fig. 3(b), thus, the outside
wall temperature of the inside tube is equals to the inside wall temperature which is equals to the uids temperature. As the inside
tube carbon dioxide ow is turbulent, so its boundary layers thickness is thinner than laminar ow, therefore, we can use the out wall
temperature to replace the uids bulk mean temperature, besides,
the inside super critical carbon dioxide pressure is particularly high,
even can be higher than 10 MPa, so its difcult to determine the
inside uids temperature, as a result, we can use thermal couples
to determine the out wall tube temperature, and use the out wall
temperature to calculate the heat transfer coefcient.
5. Experiment results
Heat transfer rate Q can be obtained by water side and carbon
dioxide side. They are calculated by Eq. (8).

Q w mw cp;w DT w
1
R1
;
hi Ai

1
do
R2
ln ;
2pkl di

1
R3
ho Ao

1
do
1
R0
ln
2pkl di pho do l
LMTD

Dt max  Dtmin
ln DDttmax
min

Then, we need to get the inside tube convection thermal resistance R1.
First, we let hot water ows in the inside tube, and cold water
ows outside tube, we call it condition 1. In condition 1, we can
calculate total thermal resistance Rcondition 1 by Eq. (1), then we
can use Gnielinski correlation (Eq. (7)) [20] to calculate hot water
side heat transfer coefcient, and calculate inside water convection
thermal resistance by Eq. (4).

Nuf

f =8Re  1000 Prf

p
1 12:7 f =8 Prf2=3  1

Second, we can dry the inside tube and let carbon dioxide ows
inside, we can calculate the total thermal resistance R. This is the
condition 2. In condition 1 and 2, we keep the inlet temperature
and ow speed of cold water constant, then the thermal resistance
of R is constant in both condition.
Finally, we use R in condition 2 minus R in condition 1 to get
the carbon dioxide side thermal resistance, after that, we can use
Eq. (4) to calculate the heat transfer coefcient of carbon dioxide.

Q c mc iin;c  iout;c

In Eq. (8), subscript w means water, and c means carbon

dioxide, i means enthalpy, cp means constant-pressure specic
heat. We use water side heat transfer rate Qw to calculate.
On dealing with the heat transfer rate, relative heat balance
error needs to be less than 15%, thus the data is considered as
believable. The heat balance can be calculated by equation 9,
where Qmax mean the max of Qc and Qw. In fact, most of the relative
heat balance is less than 15%.

Q max  Q min
Q max

Tb in Eq. (10) is used as the properties evaluated bulk mean

temperature.

Tb

T in T out
2

10

Tin means carbon dioxide inlet temperature of the test section

(out wall temperature), Tout means outlet temperature. The detail
of the R744s inlet/outlet temperature of our experiment is shown
in appendix Tables 1 and 2.
The average heat transfer coefcient h for the test section can
be calculated by Eqs. (11) and (12).

DT
1
; Rc R  R0
Q
pdi lhi

11

hi

1
R  R0 pdi l

12

4. Experiment equipment and procedure

5.1. Effects of temperature
The scheme of experiment equipment is showed in Fig. 2. It
consisted of a high pressure carbon dioxide pump, a ow meter,
some spring safety valves (which can also be called pressure valve,
it can release carbon dioxide in case the pressure is higher than
12 MPa, higher pressure may be dangerous), pre-heater, test
section, Coriolis Mass Flow meter, thermal couples, two pressure
sensors (one sensors to measure inlet uid pressure and the other
for outlet), one difference pressure sensor (to measure the pressure
loss when the uid ows through the test section).

As can be seen in Fig. 4, At a certain pressure and mass ow rate,

heat transfer coefcient changes with temperature. With the
increase of temperature, h rst increase, in the region near the
pseudo-critical point, h gets its max, when temperature is higher
than pseudo point, with temperature increase, heat transfer coefcient decrease. It can be also seen from Fig. 4, in the region near
the pseudo-critical point, gradient of the curve is higher than the
region far away from pseudo-critical point. In other words, the

T. Ding, Z. Li / International Journal of Heat and Mass Transfer 92 (2016) 665674

669

Fig. 2. The scheme of experiment equipment.

more region closer to the pseudo-point, the more violent change of

heat transfer coefcient.
In fact, on a certain pressure, at pseudo-critical point specic
heat gets its max, for example, the pressure is 8 MPa, at 35.4 C
heat transfer coefcient gets its max. Considering the measure
error, we can conclude from Fig. 5, heat transfer coefcient gets
its max at pseudo-critical point. The variation tendency of the
specic heat and heat transfer coefcient is just the same. In other
words, the heat transfer coefcient severely changeable region is
also the region where thermal property change rapidly.
The difference of super critical uid and single phase uids
appears in three aspects.
1. There is no phase change for super critical uid. When temperature is less than pseudo-critical point, the uid has high viscosity and density, uids properties are just like gas. When
temperature is higher than pseudo critical point uids properties are just like liquid.
2. Variation of thermal properties. Ordinary single phase uids
thermal properties changes slowly with temperature, so we
can regard such uid as constant property uid. For super
critical uid, at a certain pressure, when temperature is

considerably higher or lower than pseudo-point, we can also

regard such uid as constant property uid.
3. Variation of heat transfer coefcient. Thermal properties, physical dimension and mass ow rate determine the heat transfer
coefcient, so the variation tendency of heat transfer coefcient
and thermal properties are roughly analogous.
5.2. Effects of pressure
Fig. 6 shows heat transfer coefcient changes with temperature and pressure, when mass ow rate is 220 kg/h. We can
see from Fig. 6 that, besides temperature, pressure also makes
a great difference for heat transfer coefcient. At 8 MPa, hs peak
value is 16516 W/m2 K, with the increase of pressure, heat transfer decreased, at 9 MPa, 40.1 C, heat trasfer coefcients peak
value is 13,259 W/m2 K, 10 MPa, there is no obvious peak
appears, at pseudo critical point, its heat transfer coefcient is
9286 W/m2 K.
It can also be seen from Figs. 7 and 6 when the pressure is high,
the curves peak is gentle, peak comes more steeper when pressure
comes lower. In fact, the curves peak at 8 MPa is far more steeper
than peak at 10 MPa.

670

T. Ding, Z. Li / International Journal of Heat and Mass Transfer 92 (2016) 665674

Fig. 3. Test section.

16000

14000

1.5

x/xpc

h(W/m2K)

18000
Density/Density_pc
k/k_pc
Viscosity/viscosity_pc
Cp/Cp_pc
h/h_max

12000
1

10000
0.5

8000
6000
28.0

0
0.8

33.0

38.0

43.0

48.0

T/oC
Fig. 4. 8 MPa, mass ow rate is 220 kg/h, heat transfer changes with temperature.

Fig. 7 shows the specic heats variation tendency with temperature and pressure. The variation of specic heat and heat transfer
coefcient shows the same tendency.
5.3. Effects of mass ow rate
There are 3 curves in Fig. 8 (pressure is 8 MPa) and also 3 curves
in Fig. 9 (pressure is 9 MPa). They represent the experiment results
of heat transfer coefcient at different mass ow rates. In theory,

0.9

1.1

1.2

1.3

1.4

T/Tpc
Fig. 5. Comparison of the variation trends between Cp and h.

increasing mass ow rate will enhance heat transfer, experiment

results also shows the same tendency.
5.4. The comparison of different empirical formula and our experiment
results
In order to compare, we choose a typical condition, its pressure
is 8 MPa and mass ow rate is 220 kg/h. There are four results in

T. Ding, Z. Li / International Journal of Heat and Mass Transfer 92 (2016) 665674

671

5.5. Uncertain analysis

Errors always consist in any experiment. It is important to know
whether this experiment method is accurate. There are mainly two
reasons lead to error. The rst is the systematic error, which is
caused by experiment method and the second is measurement
error which is caused mainly by the accuracy of the measurement
instruments.
We can use uncertainty analysis to know measurement error.
There are mainly 3 indirect measured physical properties when
we calculate h, they are heat transfer capacity Q (J/s), LMTD (K)
and thermal resistance (Ks/J).
The uncertainty of physical properties x1  x2 can be calculated
by Eq. (13).
Fig. 6. Mass ow rate is 220 kg/h, heat transfer coefcient changes with pressure.

Z x1  x2
s s

2 
2




DZ
@lnZ
@lnZ
Dx1 2
Dx2 2

Dx1
Dx2

Z
@x1
@x2
x1  x2
x1  x2

40

13
8MPa
10MPa

35

9MPa

In Eq. (13), if x means temperature, then Z means temperature

difference. If x1 means Rtotal and x2 means R, then Z means Rco2.
The uncertainty of LMTD can be obtained by Eq. (14), here, DT
mean the measurement error of temperature difference, which is
0.14 K calculated by Eq. (13).

Cp(kJ/kg/K)

30
25
20
15

DT max  DT min


ln DDTTmax
min
s

2 
2
Df
@ ln f
@ ln f

DT
DT
f
@ DT max
@ DT min
v
!2
!2
u
DT
DT
u
DT
DT
DT min
DT max
t

DT max  DT min ln DDTTmax

DT max  DT min ln DDTTmax
min
min

f LMTD

10
5
0
20

25

30

35

40

45

50

55

60

T/oC
Fig. 7. Specic heat changes with temperature and pressure.

14
Fig. 10, Yoons empirical equation, DittusBoleter equation,
Gnielinski equation and our experiment results. It can be seen that
DB empirical equation leads to the lowest result, while Yoons
equation gets the highest result. When temperature is higher than
36 C, experiment results are less than Yoons result but around
30% higher than Gnielinski equation. However, when temperature
is higher than 33 C and less than 36 C, in other word, when temperature is near the pseudo critical point, (33 C36 C) heat transfer coefcients peak value is 25% less than Gnielinski equation.

The uncertainty of heat transfer capacity (U) can be obtained by

Eq. (15), where DT means temperature difference, and Dt means
the uncertainty of temperature difference, which is also 0.14 K.

U cp mDT
DU

s s

2 
2


 2
@ ln U
@ ln U
Dm 2
Dt
Dm
Dt

m
@m
@t
DT

Fig. 8. Heat transfer coefcient changes with mass ow rate when pressure is 8 MPa.

15

672

T. Ding, Z. Li / International Journal of Heat and Mass Transfer 92 (2016) 665674

Fig. 9. Heat transfer coefcient changes with mass ow rate when pressure is 9 MPa.

60000

Experiment

50000

Dittus-Boleter

h/Wm2K

40000

Gnielinski

Dh

Yoon
30000
20000

31.0

36.0

41.0

46.0

T/oC
Fig. 10. Comparison of different empirical formula and my experiment results
8 MPa, 220 kg/h.

The uncertainty of thermal resistance can be obtained by Eq.

(16)

DT m
Us s

2 
2
 2  2
DR
@ ln R
@ ln R
Df
DU

Df
DU

R
@f
@U
f
U

17

s
 2 

 2
DRco2 2
Ddi
Dl

l
di
Rco2

Table 3 is the result of uncertainty for each physical properties.

In fact, different condition (different temperature and pressure)
leads to different heat transfer coefcient, in the region near the
peseudo-critical point, uncertainty can be as high as 1520%, however, in the region away from the peseudo-critical point, uncertainty can be lower than 10%. It is easy to explain such
phenomenon. In the region near the peseudo-critical point, physical properties and heat transfer coefcient change violently with
temperature and pressure, a little measurement uncertainty of
temperature can lead to large uncertainty of heat transfer
coefcient.

10000
0
26.0

p di Rco2 l

5.6. A new correlation for super critical carbon dioxide

16

As for DittusBoleter equation and Gnielinski equation, they are

used to calculate the heat transfer of ordinary uid like water or
Freon in single phase. It may not t for super critical carbon dioxide, thus, there are differences between these equations and our
experiment results. In order to solve this problem, a new correlation between heat transfer coefcients and Reynolds number and
Prandtl number is developed. Where Re and Pr is dened in
Eq. (1), as is mentioned above. We use the EES (version 9.941) software to calculate thermal properties, and the results of the EES
software is based on reference [2123].

Supposing that R and Rtotal has the same uncertainty, then the
uncertainty of Rco2 can be calculated by Eq. (13).
Finally, we can obtain the total uncertainty of heat transfer
coefcient by Eq. (17), d is the diameter of the inside tube, which
is 7.50 mm, it is measured by vernier caliper, whose accuracy is
0.02 mm. L is the total length of the test section, which is 3 m
(inside tube is 3 m, outside cold water tube is 2.9 m), its accuracy
is 0.1 cm.

Table 3a
Uncertainty of heat transfer coefcient at the region far away from pseudo critical point.
Physical properties

LMTD (K)

Heat transfer
capacity (J/s)

Rtotal (Ks/J)

Rco2 (Ks/J)

Test section length (m)

Inside tube
diameter (mm)

Heat transfer
coefcient (W/m2 K)

Uncertainty

0.39%

2.51%

2.53%

7.75%

0.03%

0.27%

7.75%

Table 3b
The max uncertainty of heat transfer coefcient, when pressure is 8 MPa, temperature is 35.4 C.
Physical properties

LMTD (K)

Heat transfer
capacity (J/s)

Rtotal (Ks/J)

Rco2 (Ks/J)

Test section
length (m)

Inside tube
diameter (mm)

Heat transfer coefcient

(W/m2 K)

Uncertainty

0.58%

2.90%

2.96%

22.93%

0.03%

0.27%

22.93%

673

T. Ding, Z. Li / International Journal of Heat and Mass Transfer 92 (2016) 665674

Nu 0:028332Re0:837992 Pr0:078006
5

1:2  10 6 Re 6 4:97  10 ;

18

1:80 6 Pr 6 13:16

Liner regression is used to nd the correlation, which is Eq. (18),

in this equation, the application range of Re and Pr is also shown.
The equation ts well with our experiment results, there are 49%
of the data whose results difference calculated by empirical correlation and experiment data is less than 5%. There are 91% of all the
data whose results difference calculated by empirical correlation
and experiment data is less than 15%.

6. Conclusions
(1) The variation of heat transfer and thermal property have the
same trends. Thermal properties change with temperature
and pressure, which leads to the change of heat transfer
coefcient. In the region where thermal properties change
slowly (like ordinary single phase uid), the heat transfer
coefcient also change slowly. However, in the region thermal properties change violently (like super critical carbon
dioxide), the heat transfer coefcient also change quite a lot.
(2) In super critical region, heat transfer changes violently with
temperature, as a result, on dealing with the super critical
regions convection heat transfer problem, we cant regarded
such uid as ordinary single phase uid, however, in the
region far away from the psuedo-critical point, we can
regarded it as single phase uid.
(3) Heat tranfer coefcient changes with temperature and pressure. At a certain pressure, with increase of temperature,
heat transfer coefcient rst increase then decrease, heat
transfer coefcient gets its max at pseudo-critical point. As
for pressure, heat transfer coefcient curves peak value is
higher when pressure is higher, when the pressure is high,
curves peak is gentle, peak comes more steeper when pressure comes lower. In fact, the curves peak at 8 MPa is far
more steeper than peak at 10 MPa.
(4) As for DittusBoleter equation and Gnielinski equation, they
are used to calculate the heat transfer of ordinary uid like
water or Freon in single phase. It may not t for super critical
carbon dioxide, thus, there are differences between these
equations and our experiment results. In order to solve this
problem, a new correlation of heat transfer coefcients
based on our experiment results is developed.

Table A1
The results of R744s inlet and outlet temperature, pressure is 8 MPa.
8 MPa, 220 kg/h

8 MPa, 180 kg/h

8 MPa, 140 kg/h

Inlet (C)

Outlet (C)

Inlet (C)

Outlet (C)

Inlet (C)

Outlet (C)

56.2
53.5
50.8
48.7
46.6
43.4
40.8
39.2
37.7
36.8
36.3
35.3
34.7
34.2
32.9

36.3
36.0
35.9
35.6
35.0
35.2
34.7
34.6
34.3
34.0
33.8
32.6
30.9
29.3
26.7

58.5
56.8
54.9
52.8
49.7
47.1
43.7
40.0
37.3
36.2
35.1
34.6
33.2

35.6
35.4
35.3
35.0
34.8
34.6
34.5
34.2
33.6
33.0
30.8
30.1
25.7

57.1
55.1
53.0
48.6
45.9
42.4
40.7
38.8
37.9
37.3
36.3
35.8
34.8

35.2
34.8
34.9
34.6
34.3
34.3
34.1
33.8
33.4
33.1
32.1
31.7
28.9

Table A2
The results of R744s inlet and outlet temperature, pressure is 9 and 10 MPa.
9 MPa, 220 kg/h

9 MPa, 180 kg/h

9 MPa, 140 kg/h

10Mpa 220 kg/h

Inlet
(C)

Outlet
(C)

Inlet
(C)

Outlet
(C)

Inlet
(C)

Outlet
(C)

Inlet
(C)

Outlet
(C)

56.5
53.3
51.9
49.0
47.2
44.8
43.8
43.2
42.3
41.4
40.3
39.1
37.7

40.5
40.0
39.6
38.9
38.5
37.7
37.2
36.9
36.0
35.1
33.6
30.8
29.2

57.2
54.6
52.5
50.0
48.0
45.3
44.3
42.9
42.1
40.7
39.4
38.0

39.8
39.4
38.8
37.9
37.2
36.4
36.1
35.3
34.2
32.3
29.9
28.3

57.6
54.7
52.7
50.3
48.8
45.9
45.1
43.9
42.0
40.3
39.0

38.5
38.1
37.5
36.7
35.0
35.1
34.7
33.9
32.5
30.4
27.7

42.9
45.0
46.9
47.9
48.8
49.5
50.9
51.8
52.2
57.3
61.6
65.0

31.6
34.0
35.5
36.1
36.8
37.4
38.3
39.0
39.6
41.6
43.0
43.9

Conict of interest
None declared.
Acknowledgments
This paper is supported by National Science Foundation of
China (51376097, 51138005).
Appendix A
The detail results of R744s inlet and outlet temperature
(see Tables A1 and A2).
References
[1] J.P. Holdren, J. Ahearne, R. Balzhiser, et al., Report to the president on federal
energy research and development for the challenges of the twenty-rst
century, Presidents Committee of Advisors on Science and Technology (PCAST),
Panel on Energy Research and Development, Washington DC, 1997.
[2] X.R. Zhang, H. Yamaguchi, K. Fujima, et al., A feasibility study of CO2-based
Rankine cycle powered by solar energy, JSME Int. J. Ser. B 48 (3) (2005) 540
547.
[3] G. Lorentzen, J. Pettersen, A new, efcient and environmentally benign system
for car air-conditioning, Int. J. Refrig. 16 (1) (1993) 412.
[4] A. da Silva, E.P. Bandarra Filho, A.H.P. Antunes, Comparison of a R744 cascade
refrigeration system with R404A and R22 conventional systems for
supermarkets, Appl. Therm. Eng. 41 (2012) 3035.
[5] X.D. Qian, H. Tian, Z. Li, Z.X. Li, Entransy-dissipation-based thermal resistance
analysis and energy saving design of data center cooling system, in:
Proceedings of the 3rd International Forum on Heat Transfer, Nagasaki,
Japan, no. IFHT 2012027, November 1315, 2012.
[6] Y.W. Zheng, Z. Li, X.H. Liu, Z. Tong, R. Tu, Retrot of air conditioning system in
data center using separate heat pipe system, in: Proceedings of the 8th
International Symposium on Heating, Ventilation and Air Conditioning, Xian,
China, pp. 685694, 2013, http://dx.doi.org/10.1007/978-3-642-39581-9_67.
[7] S.J. Jeong, An experimental study on heat transfer coefcients of a CO2-lled
thermosyphon, Exp. Heat Transfer 24 (3) (2011) 257265.
[8] S.H. Yoon, J.H. Kim, Y.W. Hwang, et al., Heat transfer and pressure drop
characteristics during the in-tube cooling process of carbon dioxide in the
supercritical region, Int. J. Refrig. 26 (8) (2003) 857864.
[9] C.H. Son, S.J. Park, An experimental study on heat transfer and pressure drop
characteristics of carbon dioxide during gas cooling process in a horizontal
tube, Int. J. Refrig. 29 (4) (2006) 539546.
[10] S.M. Liao, T.S. Zhao, An experimental investigation of convection heat transfer
to supercritical carbon dioxide in miniature tubes, Int. J. Heat Mass Transfer 45
(25) (2002) 50255034.
[11] S.M. Liao, T.S. Zhao, Measurements of heat transfer coefcients from
supercritical carbon dioxide owing in horizontal mini/micro channels, J.
Heat Transfer 124 (3) (2002) 413420.
[12] X.L. Huai, S. Koyama, T.S. Zhao, An experimental study of ow and heat
transfer of supercritical carbon dioxide in multi-port mini channels under
cooling condition, Chem. Eng. Sci. 60 (12) (2005) 33373345.
[13] P.X. Jiang, Y.J. Xu, J. Lv, et al., Experimental investigation of convection heat
transfer of CO2 at super-critical pressures in vertical mini-tubes and in porous
media, Appl. Therm. Eng. 24 (8) (2004) 12551270.

674

T. Ding, Z. Li / International Journal of Heat and Mass Transfer 92 (2016) 665674

[14] S. He, P.X. Jiang, Y.J. Xu, et al., A computational study of convection heat
transfer to CO2 at supercritical pressures in a vertical mini tube, in:
ASME 2004 2nd International Conference on Microchannels and
Minichannels, American Society of Mechanical Engineers, 2004, pp. 297
304.
[15] J. Petterse, Heat transfer and pressure drop characteristics of supercritical
carbon dioxide in microchannel tubes under cooling, in: Lorentzen 4th IIR
Gustav Lorentzen Conference on National working uids, Purdue University
Press, 2000, pp. 99106.
[16] C.B. Dang, E. Hihara, In-tube cooling heat transfer of supercritical carbon
dioxide. Part 1. Experimental measurement, Int. J. Refrig. 27 (7) (2004) 736
747.
[17] J. Lv, Study of performance of cycle and heat transfer of CO2 trans-critical cycle
(PhD thesis), Tianjing University, Tianjin, 2005 (in Chinese).

[18] N. Sawant, M.S. Kim, W. Payne, et al., A study of in-tube evaporation heat
transfer of carbon dioxide, in: International Conference of Refrigeration, 2003,
pp. 18.
[19] F.P. Incropera, D.P. Dewitt, Fundamentals of heat and mass transfer, seventh
ed., John Wiley & Sons, 2011. pp. 434.
[20] V. Gnielinski, New equations for heat and mass-transfer in turbulent pipe and
channel ow, Int. Chem. Eng. 16 (2) (1976) 359368.
[21] V. Vesovic, W.A. Wakeham, G.A. Olchowy, et al., The transport properties of
carbon dioxide, J. Phys. Chem. Ref. Data 19 (3) (1990) 763808.
[22] A. Fenghour, W.A. Wakeham, V. Vesovic, The viscosity of carbon dioxide, J.
Phys. Chem. Ref. Data 27 (1) (1998) 3144.
[23] R. Span, W. Wagner, A new equation of state for carbon dioxide covering the
uid region from the triple-point temperature to 1100 K at pressures up to
800 MPa, J. Phys. Chem. Ref. Data 25 (6) (1996) 15091596.