Cold Regions Science and Technology 95 (2013) 67–73

Contents lists available at ScienceDirect

Cold Regions Science and Technology

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

Laboratory investigation of the heat transfer characteristics of a

two-phase closed thermosyphon
Mingyi Zhang a,c,⁎, Yuanming Lai a, Yuanhong Dong b, Long Jin a,b, Wansheng Pei a, Jon Harbor c
a
State Key Laboratory of Frozen Soil Engineering, Cold and Arid Regions Environmental and Engineering Research Institute, Chinese Academy of Sciences, Lanzhou, Gansu 730000, China
b
Key Laboratory of Highway Construction & Maintenance Technology in Permafrost Regions, Ministry of Transport, CCCC First Highway Consultants Co., LTD, Xi'an, Shaanxi 710065, China
c
Department of Earth, Atmospheric, and Planetary Sciences, Purdue University, West Lafayette, IN 47907, USA

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

Article history: Two-phase closed thermosyphons (TPCTs) are widely used as heat transfer devices in engineering structures in
Received 10 January 2013 permafrost regions to prevent freeze–thaw damage. In this paper, we reported on a laboratory experiment to
Accepted 8 August 2013 study the heat transfer characteristics and working state of a TPCT installed in the soil based on the typical
temperature conditions in the permafrost regions of the Qinghai–Tibet Plateau. The results indicated that
Keywords:
when the temperature of the condenser section was lower than that of the evaporator section, and the
Laboratory investigation
Heat transfer characteristic
negative temperature difference between the condenser and evaporator sections was beyond a critical temper-
Thermal semi-conduction effect ature difference, the TPCT was active; otherwise, the TPCT was inactive. The critical temperature difference
Two-phase closed thermosyphon between the condenser and evaporator sections that was required for the TPCT to begin to work was defined
Permafrost region as the “startup temperature difference”. Furthermore, we found that the efficiency of the TPCT improved linearly
with the negative temperature difference between the condenser and evaporator sections when the TPCT was
working. In this study, the startup temperature difference of the TPCT was about −0.20 °C and the total thermal
resistance was about 0.31 °C/W. These results provide a basis for future work evaluating and improving the per-
formance of TPCTs used for embankment design in permafrost regions.
© 2013 Elsevier B.V. All rights reserved.

1. Introduction
Two-phase closed thermosyphons (TPCTs) are used as heat transfer
devices in many engineering applications, including heat exchangers,
solar energy conversion systems, and civil engineering projects, because
they are simple devices that have easy operation and maintenance
(Noie, 2005). TPCTs are used in engineering projects in permafrost re-
gions in the USA, Canada and Russia (Yang et al., 2005); in China this
technology has been widely applied to protect permafrost below em-
bankments in cold regions, including the Qinghai–Tibet Railway and
Highway. Field studies of embankments indicate that TPCT can effec-
tively reduce underlying ground temperature and improve embank-
ment stability in permafrost regions (Cheng et al., 2008; Li et al., 2003;
Pan et al., 2003; Sun, 2005; Wang et al., 2005; Wu et al., 2010; Yang
et al., 2005).
The cooling effect of the TPCT results from a “thermal semi-
conduction effect” caused by vaporization and condensation of a fluid
in a closed device that has an evaporator and a condenser. In an em-
bankment application, the evaporator section is buried in the soil that
makes up the subgrade, and the condenser section is above ground in
⁎ Corresponding author at: State Key Laboratory of Frozen Soil Engineering, Cold and
Arid Regions Environmental and Engineering Research Institute, Chinese Academy of
Sciences, Lanzhou, Gansu 730000, China. Tel.: +86 931 4967296.
E-mail address: myzhang@lzb.ac.cn (M. Zhang).
the open air. When the temperature of the condenser section is lower
than that of the evaporator section, and the negative temperature differ-
ence between the condenser and evaporator sections is beyond a critical
temperature difference, the working fluid in the evaporator section
absorbs external heat energy and vaporizes. The vapor rises to the con-
denser section because of the vapor pressure gradient. In the condenser
section, the vapor is condensed on the wall, releasing latent heat, and
the liquid returns down along the wall to the evaporator section
under the influence of gravity. Otherwise, the heat transfer process
stops. Thus, in a cycle of increasing and decreasing air temperature,
when the TPCT is in a working state, heat energy from soil can be
released through the TPCT; however when the TPCT stops working,
no heat energy is transferred by the TPCT mechanism. Besides, a small
amount of heat can only be exchanged between soil and air through
conduction along the TPCT wall during the whole temperature
change cycle.
Previous studies have focused on the cooling effect of the TPCT and
its applications in cold regions engineering. Wu et al. (1996) tested con-
vection heat transfer between a TPCT and air in a laboratory setting and
showed that the heat transfer coefficient of a TPCT was related to the
air temperature and outside wind speed. Pan and Wu (2002) studied
how soil freezing expansion was restrained by TPCTs using numerical
modeling, and Pan et al. (2003)'s study of the cooling radius and the
temperature distribution of TPCT embankment in the Qinghai–Tibet
Plateau concluded that the TPCT was effective in protecting permafrost

0165-232X/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.coldregions.2013.08.006
68 M. Zhang et al. / Cold Regions Science and Technology 95 (2013) 67–73

Speeding fan Wind velocity-controlling facility

Passage for wind circulation Auto temperature
controller
Wind Temperature sensor
Model box
Cooling fan
TPCT

2.7 m Wind
Model Soil

Insulation wall Double-head

Temperature sensor
compressor
or heat flux sensor
Data logger
Desk top computer
8.0 m

Fig. 1. Schematic of the experimental equipment.

and enhancing embankment stability. Wang et al. (2005) also drew the
same conclusion by analyzing the working state and cooling effect of
TPCT embankments of the Qinghai–Tibet Highway. Wu et al. (2010)
found that combining a TPCT with insulation was a more effective
way to decrease ground temperatures, based on the data collected on
an embankment along the Qinghai–Tibet Railway.
Recently, numerical studies have been performed to predict and
evaluate the long-term thermal stability of TPCT-protected embank-
ments in permafrost regions. Liu et al. (2004) analyzed the cooling effect
of a TPCT system on a permafrost roadbed using an approach that
modeled the TPCT as a line with an approximate heat flow. Wen et al.
(2005) analyzed the temperature characteristics of the embankment
protected by the combination of both insulation and TPCTs in perma-
frost regions by utilizing the same simplification of the TPCT. Similarly,
based on the heat flow characteristics of TPCTs derived from field obser-
vations, Yang et al. (2006) used a numerical approach to study the
cooling effect of TPCTs with different inclination angles for embank-
ments along the Qinghai–Tibet Railway. Using a more complex model-
ing approach, Zhang et al. (2011) studied the thermal characteristics of
a TPCT-protected embankment along the Qinghai–Tibet Railway using
a coupled heat transfer model of air–TPCT–soil for the TPCT embank-
ment based on heat transfer theories.
Generally speaking, the cooling capacity of a TPCT and its general ap-
plication as an approach to enhance embankment stability have been
the primary foci of research to date. However what is less well
understood are the heat transfer processes and performance character-
istics for a TPCT used to reduce ground temperatures in cold regions. In
the work reported here, we studied the heat exchange characteristics of
a TPCT by analyzing the temperature and heat transfer rate changes of a
TPCT model in a laboratory setting. Better understanding of the perfor-
mance of TPCTs will be helpful in the future design of TPCT embank-
ments in permafrost regions.
2. Experimental design
The experimental work took advantage of a custom-designed cold
regions engineering laboratory that was set up for scale model simula-
tions. The experimental equipment consisted of four parts: a model box,
a temperature controlling system, a ventilation system and a data acqui-
sition system (Fig. 1). The model box was insulated and has inner di-
mensions of 8.0 m × 1.84 m × 2.7 m (length, width and height). The
temperature control system was composed of a double-head compres-
sor and an automatic temperature controller with a temperature sensor.
The ventilation system was composed of cooling fans, speeding fans,
wind velocity-controlling facilities and a passage for wind circulation.
The wind direction was parallel to the longitudinal direction of the
model box. The data acquisition system included temperature sensors,
heat flux sensors, a data logger and a desk top computer.
The experimental model for the TPCT is shown in Fig. 2. The cross-
section of the model was 2.5 m × 1.0 m along the longitudinal direction

TPCT

Condenser section

1
0.06 m

2
0.253 m

9 Temperature sensor

Insulation board I
3 8 10 15 16
Evaporator section
1.0 m

4 II
0.267 m

Temperature sensor 7 11 14 17
and heat flux sensor
III
5 6 12 13 18

0.2 m 0.15 m
0.1 m 2.5 m

Fig. 2. Schematic of the experimental model for TPCT.

 M. Zhang et al. / Cold Regions Science and Technology 95 (2013) 67–73 69

12 30 18
Condenser section Heat transfer rate Temperature difference

Temperature difference Δ T (°C)

8
Surface temperature T (°C)

Heat transfer rate Q (W)

20 12
4

0 10 6

-4
0 0

-8

Evaporator section -10 -6

-12 0 216 432 648 864 1080
0 216 432 648 864 1080
Time(h) Time(h)

Fig. 5. Heat transfer rate of the TCPT vs. temperature difference between the condenser
Fig. 3. Surface temperature change of the TPCT with time.
and evaporator sections.

of the model box. The length of the TPCT was 1.525 m, the outer diam-
after the first temperature cycle. The condenser section reached a
eter was 0.051 m, and the working fluid was ammonia. In this study, the
highest temperature of 8 °C, which was the highest air temperature
evaporator section was taken as half of the TPCT, and buried in the soil
in the model box, and the condenser section had a lowest tempera-
layer. The condenser section was exposed to the air. The soil layer was
ture that did not get lower than − 12 °C (Fig. 3), even though the
enclosed by insulation boards. To study the heat transfer characteristics
lowest air temperature in the model box was − 16 °C. This differ-
of the TPCT, the temperature and heat flux were tested in the TPCT
ence is due to the heat flux from the evaporator to the condenser.
model. The model dimensions and the locations of the temperature sen-
In the last three cycles the lowest temperature reached in the con-
sors and heat flux sensors are shown in Fig. 2.
denser section was dropping slowly, and we predicted that this
Prior to the starting of the experiment, the installed model was kept
would continue as the model slowly trended towards an equilibri-
at indoor temperature (at approximately 16 °C) for 72 h to reach a ther-
um condition.
mally stable state. Experimental conditions were then applied. Based on
In the evaporator section, we defined the surface temperature as the
air temperature records on the Qinghai–Tibet Plateau and related data
mean temperature value of sensors 3 and 4 (Fig. 2). The surface temper-
(Cheng et al., 2003; Lai et al., 2009), the ambient temperature in the
ature of the evaporator also changed over time in phase with the surface
model box was controlled by a sinusoidal function:
temperature of the condenser section, and the time pattern of the tem-
perature change became relatively stable after the fourth cycle. In this
T ¼ −4:0−12:0 sinð2πt=216Þ ð1Þ stable mode the surface temperature of the evaporator section stayed
below 0 °C throughout the entire temperature change cycle (Fig. 3).
where t is time in hour; the mean periodic temperature was −4.0 °C, The mean periodic surface temperature of the evaporator section
the temperature amplitude every cycle was 12.0 °C, and cycle period was higher than that of the condenser in the first two cycles (Fig. 4)
was 216 h (9 days). The wind velocity was 2.8 m/s across the top because of the influence of the initial temperature of the soil around
of the model. The test conditions were repeated periodically and five the evaporator (the soil temperature at the start of the experiment
cycles were completed with test data collected at intervals of 20 min. was about 16 °C); however, from the third cycle, the mean periodic
surface temperature of the evaporator section was lower than that
3. Experimental results and analysis of the condenser section. By the fifth cycle, the mean periodic surface
temperature of the evaporator section was − 2.75 °C, which was
During the experiment the surface temperature of the condens- 0.74 °C lower than that of the condenser section, and 4.85 °C lower
er section of the TPCT (sensor 1; Fig. 2) changed periodically with
the ambient air temperature, and had a relatively stable pattern
20

3
Mean periodic surface temperature Tp (°C)

Q = -3.24Δ T - 0.66
Heat transfer rate Q(W)

15
2 R2 = 0.8997
Condenser section Evaporator section
1
10

0
5
-1

-2 0
0 -1 -2 -3 -4 -5
Temperature difference Δ T(°C)
-3
0 1 2 3 4 5 6
Cycle Fig. 6. Variation of the heat transfer rate of the TCPT with temperature difference between
the condenser and evaporator sections when the TPCT was in working state in the fifth
Fig. 4. Mean periodic surface temperature change of the TPCT. cycle.
70 M. Zhang et al. / Cold Regions Science and Technology 95 (2013) 67–73

than the mean periodic surface temperature of the evaporator in the
first cycle. This shows that the TPCT acted as an effective conduit of
heat out of the soil during the initial two cycles when the soil was ini-
tially much warmer than the average air temperature in the model
box. After the third cycle, the TPCT maintained an evaporator section
temperature that did not go above freezing despite air temperature
in the model box that increased to 8 °C.
The mean outside surface heat flux at the evaporator section of the
TCPT is the mean heat flux value determined from sensors I, II and III
(Fig. 2). Therefore, The heat transfer rate of the TCPT can be calculated:
Q ¼ qe Ae ð2Þ
where Q is the heat transfer rate of the TCPT, qe is the mean outside
surface heat flux at the evaporator of the TCPT, Ae is the outside surface
area of the evaporator of the TCPT.
The heat transfer rate of the TCPT and the temperature differ-
ence between the condenser and evaporator sections are shown in

a) First cycle
0 0

-0.1 -0.1

-0.2 -0.2

-0.3 -0.3
Y/m

Y/m
-0.4 -0.4

-0.5 -0.5

-0.6 -0.6

-0.7 -0.7

-0.8 -0.8
0 0.1 0.2 0.3 0.4 0.5 0 0.1 0.2 0.3 0.4 0.5
X/m X/m
Lowest ambient temperature Highest ambient temperature

b) Second cycle
0 0

-0.1 -0.1

-0.2 -0.2

-0.3 -0.3
Y/m

Y/m

-0.4 -0.4

-0.5 -0.5

-0.6 -0.6

-0.7 -0.7

-0.8 -0.8
0 0.1 0.2 0.3 0.4 0.5 0 0.1 0.2 0.3 0.4 0.5
X/m X/m
Lowest ambient temperature Highest ambient temperature

Fig. 7. Temperature distributions of the model at the lowest and highest ambient temperatures in different cycles.
 M. Zhang et al. / Cold Regions Science and Technology 95 (2013) 67–73 71

c) Third cycle
0 0

-0.1 -0.1

-0.2 -0.2

-0.3 -0.3
Y/m

Y/m
-0.4 -0.4

-0.5 -0.5

-0.6 -0.6

-0.7 -0.7

-0.8 -0.8
0 0.1 0.2 0.3 0.4 0.5 0 0.1 0.2 0.3 0.4 0.5
X/m X/m
Lowest ambient temperature Highest ambient temperature

d) Fourth cycle
0 0

-0.1 -0.1

-0.2 -0.2

-0.3 -0.3
Y/m

Y/m

-0.4 -0.4

-0.5 -0.5

-0.6 -0.6

-0.7 -0.7

-0.8 -0.8
0 0.1 0.2 0.3 0.4 0.5 0 0.1 0.2 0.3 0.4 0.5
X/m X/m
Lowest ambient temperature Highest ambient temperature
Fig. 7 (continued).

Fig. 5. When the temperature difference was negative (i.e. the sur-
face temperature of the condenser section was lower than that of
the evaporator section), the heat transfer rate of the TPCT was
taken as positive value, and increased with the negative tempera-
ture difference. When the negative temperature difference was at
its maximum, the heat transfer rate was beyond 10 W every cycle.
However when the temperature difference was positive (i.e. the
temperature of the condenser section was higher than that of the evap-
orator section), the heat transfer rate was close to zero. Furthermore,
the negative temperature difference was far smaller than the positive
temperature difference.
The heat transfer rate of the TCPT varied as a linear function of the neg-
ative temperature difference between the condenser and evaporator sec-
tions (Fig. 6), and in the fifth cycle a regression equation fit to the heat
transfer rate as a function of the negative temperature difference is given by
2
Q ¼ −3:24ΔT−0:66 R ¼ 0:8997 ð3Þ
72 M. Zhang et al. / Cold Regions Science and Technology 95 (2013) 67–73
0 0
-0.1
-0.2
-0.3
Y/m
-0.4
-0.5
-0.6
-0.7
-0.8
0 0.1 0.2 0.3 0.4
X/m
Lowest ambient temperature
Fig. 7 (continued).
where Q is the heat transfer rate of the TCPT in W and ΔT is the
negative temperature difference between condenser and evaporator
sections in °C.
For the TPCT to begin to work, a critical temperature difference
between the condenser and evaporator sections is required, and
this is defined as the “startup temperature difference”. Therefore,
Eq. (3) can be expressed as (Shiraishi et al., 1981; Tao, 2006; Zhang
et al., 2011):
1 The cooling effect of the TPCT can be seen from the changing tem-
Q ¼− ðΔT−ΔT 0 Þ
Rt
where Rt is the total thermal resistance of the TCPT, ΔT0 is defined as
startup temperature difference of the TCPT.
From Eqs. (3) and (4), the negative temperature difference between
the condenser and evaporator sections was about −0.20 °C when the
heat transfer rate of the TCPT was zero, namely, the startup temperature
difference of the TCPT was about −0.20 °C. Fig. 6 and Eq. (3) also show
that the heat transfer rate increased with the negative temperature dif-
ference, as given by the slope of the regression line: 3.24 W/°C. Based on
Eq. (4), the total thermal resistance of the TCPT can be taken as 0.31 °C/
W. The startup temperature difference and total thermal resistance are
two key measures of the performance of a TPCT; the startup tempera-
ture difference is the threshold at which the TPCT starts working, and
the total thermal resistance can be regarded as a measure of the heat
transfer efficiency of a TPCT.
As described previously, in a cycle of increasing and decreasing tem-
perature, when the temperature of the condenser section was lower
than that of the evaporator section, and the negative temperature dif-
ference between the condenser and evaporator sections was beyond
the startup temperature difference, the TPCT was in a working state
and the heat energy of soil can be released through the TPCT; otherwise,
the TPCT stopped working and no heat energy was transferred by
the TPCT mechanism. Besides, a small amount of heat can only be
e) Fifth cycle
-0.1
-0.2
-0.3
Y/m
-0.4
-0.5
-0.6
-0.7
-0.8
0.5 0 0.1 0.2 0.3 0.4 0.5
X/m
Highest ambient temperature
exchanged between soil and air through conduction along the TPCT
wall during the whole temperature change cycle. The results of this ex-
periment showed that the heat release by the TPCT in its working state
was far larger than the heat absorption (Fig. 5). The net heat energy out-
put of the TPCT resulted in a cooling of the soil, so that the mean period-
ic surface temperature of the evaporator section became lower than
that of the condenser section from the third cycle onwards (Fig. 4)
and the negative temperature difference was clearly smaller than the
positive temperature difference (Fig. 5).
ð4Þ
perature distributions of the soil around the TPCT in Fig. 7 (a–e). Soil iso-
therms were almost parallel to the TPCT at the lowest ambient
temperatures in each cycle, reflecting a large heat transfer efficiency
and a strong heat flux evenly distributed on the TPCT surface. At the
highest ambient temperatures in each cycle, the isotherms were not
parallel to the TPCT surface because the TPCT was in an inactive state.
The “U-shape” pattern of the isotherms in the upper part of the evapo-
rator indicated that heat conduction along the TPCT wall had a slight
warming effect on the soil near the surface. As expected, over the course
of the experiment, the area with temperatures below 0 °C progressively
increased, and was largest during the time when the TCPT was working
during the fifth cycle. This illustrated how the TCPT was working to pro-
gressively bring the temperature of the entire soil volume down to
values below freezing, which is important in view of preventing the
thawing of permafrost below an embankment.
The “thermal semi-conduction effect” and the general cooling effect
of the TPCT observed in this laboratory experiment is consistent with
the results of field experiments (Pan et al., 2003; Wu et al., 2010) and
numerical simulations (Wen et al., 2005; Zhang et al., 2011). Thereby,
installing TPCTs into embankments in permafrost regions can cool
down the underlying permafrost. Furthermore, it can keep the embank-
ment soil in a frozen state during the whole freeze–thaw cycle, and this
helps to protect the underlying permafrost and ensures the stability of
the embankment.
 M. Zhang et al. / Cold Regions Science and Technology 95 (2013) 67–73
4. Conclusions
A laboratory study was designed to examine the thermal behavior
of a TPCT installed in soil, based on typical temperature conditions in
the permafrost regions of the Qinghai–Tibet Plateau. The experimental
results show that:
1) The TPCT had a significant thermal semi-conduction effect and a
startup temperature difference of about − 0.20 °C. When the
temperature of the condenser section was lower than that of the
evaporator section, and the negative temperature difference be-
tween the condenser and evaporator sections was beyond the
startup temperature difference, the TPCT was active; otherwise,
the TPCT was inactive.
2) When the TPCT was in an active state, its working efficiency was
linearly related to the negative temperature difference between
the condenser and evaporator sections and the total thermal re-
sistance was about 0.31 °C/W.
These experimental results demonstrated the effectiveness of the
TPCT as a method for reducing ground temperatures in permafrost re-
gions. Future work will compare TPCTs with other heat control devices,
and optimize TPCT design to improve their heat transfer performances.
Acknowledgments
This research was supported by the Knowledge Innovation Program
of the Chinese Academy of Sciences (Grant No. KZCX2-EW-QN301),
the CAS Action-Plan for West Development (Grant No. KZCX2-XB3-19),
the National Key Basic Research Program of China (973 Program Grant
No. 2012CB026102), the National Natural Science Foundation of China
(Grant Nos. 41171060, 41101068, 41230630), the Fund of the State
Key Laboratory of Frozen Soil Engineering (Grant No. SKLFSE-ZY-03),
and Youth Innovation Promotion Association, CAS.
73
References
Cheng, G.D., Jiang, H., Wang, K.L., Wu, Q.B., 2003. Thawing index and freezing index on the
embankment surface in permafrost regions. J. Glaciol. Geocryol. 25 (6), 603–607.
Cheng, G.D., Sun, Z.Z., Niu, F.J., 2008. Application of the roadbed cooling approach in
Qinghai–Tibet railway engineering. Cold Reg. Sci. Technol. 53 (3), 241–258.
Lai, Y.M., Zhang, M.Y., Li, S.Y., 2009. Theory and Application of Cold Regions Engineering.
Science Press, Beijing, China.
Li, Y.Q., Han, L.W., Cui, L., Jia, H.F., 2003. Analysis on measurement result of probes in
Fenghuoshan area of Qinghai–Tibet Plateau. Chin. J. Rock Mech. Eng. 22 (Suppl. 2),
2669–2672.
Liu, F.X., Liu, J.K., Guo, D.H., Zhang, L.X., 2004. Numerical analysis of thermosyphon in
protecting the roadbed in permafrost. Chin. J. Civ. Eng. 37 (9), 41–47.
Noie, S.H., 2005. Heat transfer characteristics of a two-phase closed thermosyphon. Appl.
Therm. Eng. 25, 495–506.
Pan, Y., Wu, C.Z., 2002. Numerical investigations and engineering applications on freezing
expansion of soil restrained two-phase closed thermosyphons. Int. J. Therm. Sci. 41,
341–347.
Pan, W.D., Zhao, S.C., Xu, W.Z., Yu, S.S., Ma, W.D., 2003. Application of thermal probe to
enhance thermal stability of roadbed in plateau permafrost areas. J. Glaciol. Geocryol.
25 (4), 433–438.
Shiraishi, M., Kikuchi, K., Yamanishi, T., 1981. Investigation of heat transfer characteristics
of a two-phase closed thermosyphon. J. Heat Recovery Syst. 1 (4), 287–297.
Sun, Y.F., 2005. Permafrost engineering in the Qinghai–Tibet Railway: research and
practice. J. Glaciol. Geocryol. 27 (2), 153–162.
Tao, W.Q., 2006. Heat Transfer. Northwestern Polytechnical University Press, Xi'an,
China.
Wang, S.J., Huang, X.M., Chen, J.B., Zhang, J.Z., 2005. Research on frozen soil subgrade
cooling by non-power heat pipe. J. Highway and Transp. Res. and Dev. 22 (3),
1–4.
Wen, Z., Sheng, Y., Ma, W., Qi, J.L., Wu, J.C., 2005. Analysis on effect of permafrost protec-
tion by two-phase closed thermosyphon and insulation jointly in permafrost regions.
Cold Reg. Sci. Technol. 43 (3), 150–163.
Wu, Q.B., Liang, S.Y., Gao, X.W., 1996. Research of convection heat transfer between
thermalpile and air. J. Glaciol. Geocryol. 18 (1), 37–42.
Wu, J.J., Wei, M., Sun, Z.Z., Wen, Z., 2010. In-situ study on cooling effect of the two-phase
closed thermosyphon and insulation combinational embankment of the Qinghai–
Tibet Railway. Cold Reg. Sci. Technol. 60 (3), 234–244.
Yang, Y.P., Wei, Q.C., Zhou, S.H., Zhang, L.X., 2005. Thermosyphon technology and its
application in permafrost. Chin. J. Geotech. Eng. 27 (6), 698–706.
Yang, Y.P., Zhou, S.H., Wei, Q.C., 2006. Numerical analysis of proper thermosyphon
inclination angle used in permafrost embankment. Chin. Railw. Sci. 27 (3), 1–7.
Zhang, M.Y., Lai, Y.M., Zhang, J.M., Sun, Z.Z., 2011. Numerical study on cooling character-
istics of two-phase closed thermosyphon embankment in permafrost regions. Cold
Reg. Sci. Technol. 65 (2), 203–210.