Renewable Energy 177 (2021) 397e410

Contents lists available at ScienceDirect

Renewable Energy
journal homepage: www.elsevier.com/locate/renene

Heat transfer rate characteristics of two-phase closed thermosyphon

heat exchanger
Wei Song a, b, *, Changjin Zheng a, Jiaming Yang a, c
a
School of Civil and Engineering, North China University of Technology, Beijing, 100144, PR China
b
Key Laboratory of Shale Gas and Geoengineering, Institute of Geology and Geophysics, Chinese Academy of Sciences, Beijing, 100029, PR China
c
School of Architecture, Harbin Institute of Technology, Harbin, 150090, PR China

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

Article history: Two-phase closed thermosyphon heat exchangers (TPCTs) have considerable potential for use in ground
Received 10 December 2020 source heat pump systems. This paper proposes a new metal-polyethylene TPCT that uses water as the
Received in revised form working fluid and is composed of a galvanized steel pipe and a polyethylene pipe. The heat transfer rate
4 May 2021
characteristics of the metal-polyethylene TPCT were studied using a constant-temperature water bath
Accepted 24 May 2021
Available online 28 May 2021
test bench. The experimental results show that the metal-polyethylene TPCT has the best heat transfer
rate when the vacuum degree and filling ratio are 1.00 kPa and 12.5%, respectively. The average heat
transfer rate increased as the temperature of the heat source and the flow rate of the cooling water
Keywords:
Ground source heat pump
increased. Finally, a sandbox test bench was built, which provided seepage and non-seepage experi-
Two-phase closed thermosyphon heat mental conditions for the metal-polyethylene TPCT. In the sandbox experiment, the heat transfer rate
exchangers capacity of the metal-polyethylene TPCT was initially studied and compared with that of a single U-tube
Sandbox test bench heat exchanger, and the heat transfer rates were similar. When the evaporation section length was 0.5 m,
Seepage the heat transfer rates were 34.89 W/m and 111.75 W/m under non-seepage and seepage conditions,
Heat transfer rate characteristics respectively. The experimental results show that the metal-polyethylene TPCT is feasible as a new type of
ground heat exchanger.
© 2021 Elsevier Ltd. All rights reserved.

1. Introduction When the first oil crisis broke out in 1973, the price advantage of
oil disappeared. In order to find suitable substitutes for building
Building energy consumption has considerable potential for heating, some scholars began to study the GSHP system [4]. In 1979,
realizing energy savings and emission reduction. The development Bose proposed a new ground heat exchanger called the standing
and use of renewable energy can significantly reduce the energy column well (SCW), which is more efficient than a U-tube [5e7].
consumption of built structures and reduce carbon emissions. Then, in 1981 and 1983, Braud improved the operation mode and
Ground source heat pump (GSHP) systems using shallow structure of the well with regard to heat transfer rate characteris-
geothermal energy are now becoming more widely used. The his- tics and thermal resistance [8,9]. In 1986, Tan proposed the first real
torical development process of borehole heat exchangers is shown SCW [10]. Subsequently, Orio not only further improved the prac-
in Fig. 1. In 1912, Zoelly proposed a GSHP with a U-tube as a ground tical engineering of the SCW [11e13], but also studied the eco-
heat exchanger [1,2]. This type of GSHP system was successfully nomics of SCWs and evaluated their reliability [14,15]. In 1992, in
installed and used in downtown Lambert, Oregon, USA in 1946 [3]. order to alleviate the reduction of system efficiency caused by the
Although it delivered excellent performance, research in this area thermal transfixion of SCWs, the pumping and recharging well
was almost stagnant because the price of fossil fuels was low at that (PRW) was proposed by Sorensen, and preliminary tests were
time and GSHP installations required relatively high levels of conducted [16]. In 2000, to enhance the heat transfer rate near the
investment. walls of the PRW, the forced external circulation standing column
well (FECSCW) was proposed and was first put into use in Beijing,
China [17]. Nowadays, in order to prevent the groundwater level
* Corresponding author. School of Civil and Engineering, North China University
from falling, some local governments have suspended the use of
of Technology, Beijing, 100144, PR China. groundwater heat pump systems. Therefore, a new generation of
E-mail address: stillwater2013@163.com (W. Song). heat exchangers is necessary. Two-phase closed thermosyphon

https://doi.org/10.1016/j.renene.2021.05.147
0960-1481/© 2021 Elsevier Ltd. All rights reserved.
W. Song, C. Zheng and J. Yang Renewable Energy 177 (2021) 397e410

Fig. 1. Development process of borehole ground heat exchangers.

heat exchangers (TPCTs) have the advantages of low levels of In China, owing to the large size of many buildings, GSHPs have
pollution, simple structure, high thermal conductivity, and only large-scale buried pipe groups. If TPCTs are applied to GSHPs, the
require a small amount of power when extracting geothermal en- condensation section of the TPCTs should be equipped with a water
ergy [18,19]. Vertical TPCTs based on the thermosyphon principle jacket. At present, the typical TPCT tube working fluids are CO2 and
were studied in 1997 [20] and successfully installed and used in NH3. To ensure normal operation, the pressure in the tube is
Austria in 2002 [21]. maintained at 4.7 MPa [29] and 0.6 MPa [30], respectively. When
In 2002, Rieberer successfully used CO2 as the working fluid of a damaged, it not only pollutes the environment but also threatens
TPCT for a GSHP [22]. In 2004 and 2005, Kruse proved that the the safety of the personnel. In addition, the evaporation section of
efficiency of GSHPs using CO2-TPCTs is higher than that of GSHPs the TPCT should not be exposed to the temperature change layer in
that use traditional U-bend borehole heat exchangers [23e25]. In which the underground temperature is obviously affected by the
2010, Acuna determined that the average heat transfer rate of a surface air temperature, and a certain length of an insulation sec-
CO2-TPCT was 71 W/m in a 65 m deep vertical hole, which is tion should be provided. Therefore, this paper proposes the com-
significantly higher than the 30 W/m of a traditional U-tube. bined use of galvanized metal pipes and polyethylene (PE) pipes,
Moreover, it was determined that CO2 is one of the most feasible and it uses water as the TPCT working fluid. A TPCT with this
working fluids for TPCTs [26]. Subsequently, researchers focused on structure uses a metal with higher thermal conductivity in the
the theoretical aspects. In 2014 and 2017, Ebeling used a quasi- evaporation part and PE material with higher heat resistance in the
dynamic model to simulate the heat transfer rate between the adiabatic section. This not only ensures the safe operation of the
TPCT and the surrounding soil, and compared it with experimental system but also reduces the difficulty and cost of production. This
data [27,28]. Lim compared the energy consumption and coefficient structure is suitable for large underground pipeline groups and can
of performance (COP) of the three GSHPs using numerical simula- reduce pump energy consumption, thereby considerably improving
tions, and pointed out that the energy consumption of GSHPs with system efficiency [28,31]. In this paper, through a constant-
TPCTs was reduced by 10.3% and 21.1% compared with that of direct temperature water bath experiment, we explore the influence of
expansion GSHPs and secondary loop GSHPs, respectively [29]. In tube material on the condensation section and evaporation section
2020, Wang used ammonia (NH3) as the working fluid of a TPCT, construction; cooling water flow; water bath temperature and
and through simulation, found that when the working fluid pool vacuum degree and working fluid filling ratio on the heat transfer
depth is less than 10 m, a 54 m deep NH3-TPCT has good heat characteristics of the TPCT, and find the best vacuum degree and
transfer performance [30]. However, the above-mentioned studies filling ratio. Using a sandbox experiment, the heat transfer char-
all rely on the direct connection of the condensing section of the acteristics of the TPCT and a single U-tube heat exchanger under the
TPCT with the evaporator of the GSHP unit, and cannot be applied condition of seepage flow or without seepage flow are compared
to large-scale buried pipe groups. and analyzed. This work provides an experimental basis for the

398
W. Song, C. Zheng and J. Yang Renewable Energy 177 (2021) 397e410

application of water as the working fluid for TPCT in geothermal of a 1 m long PE pipe with an inner/outer diameter of 28/34 mm.
energy. The M-PE TPCT consists of evaporation and condensation sections
made of 0.5 m long galvanized steel pipes with an inner/outer
2. Materials and methods diameter of 28/34 mm. The adiabatic section is made of a 1 m long
PE pipe with an inner/outer diameter of 28/34 mm. The conden-
2.1. Composition of TPCT sation sections of both TPCTs have a 0.5 m long cooling water jacket
that surrounds the pipe with an inner/outer diameter of 51.4/
The TPCT design is shown in Fig. 2. Both a PE TPCT and a metal- 63 mm. Cooling water is passed into the cooling water jacket to
PE TPCT (M-PE TPCT) were investigated, both of which had a total exchange heat with the condensation section of the TPCTs.
length of 2 m and used water as the working fluid. The difference
lies in the materials of the evaporation and condensation sections. 2.2. Experimental setup and experiment procedure
The PE TPCT is made of PE pipes; the evaporation and condensation
sections are made of 0.5 m long PE pipes with an inner/outer The experimental setup is shown in Fig. 3, which is mainly
diameter of 28/31 mm, respectively. The adiabatic section is made composed of a constant-temperature water bath test bench and a

Fig. 2. Proposed two-phase closed thermosyphon heat exchangers (TPCT).

399
W. Song, C. Zheng and J. Yang Renewable Energy 177 (2021) 397e410

Fig. 3. Experimental setup for thermal performance test of TPCT.

sandbox test bench. Both can be switched by valves. The parame- temperature water chamber controls the constant-temperature
ters of the components are listed in Table 1, the vacuum meter is boundary conditions of the bottom boundary of the main body of
used to detect the pressure in the tube during the TPCT fabrication the sandbox. The flow rate signal is linearly transformed into a
process and is installed on the top of the TPCP (200 cm in Fig. 2). 4e20 mA direct current (DC) signal. Eleven temperature sensors
When the vacuum pump stops working, we consider the top (T0cm, T20cm … T200cm) are distributed on the outer wall of the M-PE
pressure to be the pressure in the tube. After that, the vacuum TPCT from bottom to top, as shown in Fig. 2. Other temperature
meter is removed and the TPCP is sealed. The parameters of the measurement points measure the inlet/outlet cooling water tem-
sandbox test bench are listed in Table 2. The sandbox test bench perature (Tcooling,in and Tcooling,out), the temperature of the water bath
was used to simulate the soil environment. The entire experimental (Tconst,in), and the inlet seepage water temperature (Tse,in), as shown
setup is wrapped with a thermal insulation material, which effec- in Fig. 3. The temperature sensors are T-type copper-constantan
tively prevents external factors from interfering with the experi- thermocouples. Data were collected using an Agilent 34970A data
ment. The sandbox test bench is divided into four parts: the inlet logger.
chamber, the sandbox main body, the outlet chamber, and the
bottom constant-temperature water chamber. The inlet and outlet 2.3. Description of test conditions
chambers are used so that the seepage is more evenly distributed
when simulating groundwater seepage; the main body of the The heat transfer rate (q) of the TPCT was studied in a constant-
sandbox simulates the heat transfer conditions of the TPCT under temperature water bath. In this experiment, the material of the
certain geological conditions; and the bottom constant- TPCT, the flow rate of cooling water (Q), the temperature of the

Table 1
Parameters of components.

Instruments Technical Parameters Manufacturer

Temperature Sensor T-type coppereconstantan thermocouple; Precision is ±0.10
C KAYE (X0258)

Turbine flowmeter (Cooling water) Precision is 0.50%, range is 0.060e0.600 m3/h SENLOD (SLDWY-Y6C/0.5)
Turbine flowmeter (Constant-temperature water) Precision is 0.50%, range is 0.150e1.500 m3/h SENLOD (SLDWY-Y10C/0.5)
Turbine flowmeter (Seepage water) Precision is 0.50%, range is 0.150e1.500 m3/h SENLOD (SLDWY-Y10C/0.5)
Vacuum meter Precision is 0.25%, range is 0.100e0 MPa SYCIF (YBe150B)
Data acquisition instrument Agilent 34970 A Agilent Technologies

400
W. Song, C. Zheng and J. Yang Renewable Energy 177 (2021) 397e410

Table 2
Parameters of sandbox test bench.

Sandbox test bench

Volume 1.00  0.60  1.25 m3 Sand density 1540 kg/m3

Sand porosity 36.66% Permeability coefficient 0.015 cm/s

Thermal conductivity 1.53 W/(m$
C) Thermal insulation material 0.043 W/(m$
C)

water bath (Tconst,in), the vacuum degree (p), and the filling ratio of 2.4. Data processing
the working fluid (V) are variables. There are 9 groups of working
conditions, as shown in Table 3. Condition #1 was used as the The heat transfer rate of the TPCT, calculated by Eq. (1), is
control group for the other experiments. Initial conditions #1 considered as the heat that the cooling water absorbed in the
included V ¼ 12.5%, p ¼ 2.00 kPa when the environment temper- condensation section. Here, q is the heat transfer rate of the TPCT
ature is approximately 23
C, Q is 0.2 m3/h, and Tconst,in is 31
C. After (W); Qave is average flow rate of the cooling water in the conden-
completing the above experiments, the best parameters of the TPCT sation section (m3/h); and Tcooling,out and Tcooling,in are the outlet/
were found and applied in the sandbox test bed to study the heat inlet temperatures of the cooling water, respectively (
C).
transfer rate of the TPCT and U-tubes. The data of the U-tube test
are set as reference data to verify the heat transfer rate effect of the cwater rwater Qave

q¼ Tcooling;out  Tcooling;in (1)
TPCT. See Table 4 for specific working conditions. 3600
The sandbox test is mainly divided into two working conditions,
The internal heat transfer coefficient (HTC) in the evaporation
namely non-seepage and seepage conditions. Under non-seepage
section and condensation section can be calculated using Eqs. (2)
conditions, the soil is heated from the bottom through the bot-
and (3) [32], respectively. In Eqs. (2) and (3), he is the internal
tom constant-temperature water chamber, and the test starts after
HTC of the evaporation section (W/(m2$
C)); hc is the internal HTC
the soil reaches the preset initial conditions. Under seepage con-
of the condensation section (W/(m2$
C)); qave is the average heat
ditions, the bottom constant-temperature water chamber is used
transfer rate (W); de/dc are the inner diameters of the evaporation/
for the constant-temperature boundary condition. Groundwater
condensation sections, respectively (m); le/lc are the lengths of the
seepage flows from the inlet chamber through the sandbox main
evaporation/condensation sections, respectively (m); finally, Tave,e,
body and exits the outlet chamber. The test starts when the soil
Tave,a, and Tave,c are the average temperatures of the outer wall of the
reaches the preset initial conditions.
TPCT in the evaporator, adiabatic, and condensation sections,
The cold and hot water used in the experiment are prepared by
respectively (
C).
the equipment and stored in different constant-temperature water
tanks. The flow rate of the constant-temperature water bath was
qave
maintained at 0.3 m3/h, and the temperature control accuracy was he ¼   (2)
pde le Tave;e  Tave;a
within ±0.2
C in all the conditions. The cooling water was at 10
C,
and the temperature control accuracy was within ±0.2
C. The
sandbox has a temperature gradient of approximately 3
C, qave
hc ¼   (3)
26 ± 0.5
C at the bottom and 23 ± 0.5
C at the top, and it has a pdc lc Tave;a  Tave;c
percolation speed of 8.7 m/d. During the experiment, the temper-
ature and flow rate data collection interval was 5 s. For the heat transfer rate data, the errors in each test depend on
test conditions such as flow rate and temperature [33]. All

Table 3
Constant temperature water bath test conditions.

Working condition Material Cooling water temperature Cooling water flow rate (m3/h) Water bath temperature Vacuum degree Filling ratio (%)
number (
C) (
C) (kPa)

#1 M-PE 10 0.2 31 2 12.5

TPCT
#2 PE TPCT 10 0.2 31 2 12.5
#3 M-PE 10 0.07 31 2 12.5
TPCT 0.1
0.15
0.2
0.25
0.3
#4 M-PE 10 0.2 26 2 12.5
TPCT
#5 M-PE 10 0.2 36 2 12.5
TPCT
#6 M-PE 10 0.1 31 1 12.5
TPCT
#7 M-PE 10 0.3 31 3 12.5
TPCT
#8 M-PE 10 0.2 31 2 25
TPCT
#9 M-PE 10 0.2 31 2 37.5
TPCT

401
W. Song, C. Zheng and J. Yang Renewable Energy 177 (2021) 397e410

Table 4
Sandbox test conditions.

Working condition number Material Flow rate (m3/h) Water bath temperature (
C) Vacuum degree (kPa) Filling ratio (%) Seepage conditions Seepage velocity (m/d)

#10 M-PE TPCT 0.2 26 0.1 12.5 Non-seepage e

#11 U-tube 0.2 26 e e Non-seepage e
#12 M-PE TPCT 0.2 26 0.1 12.5 Seepage 8.7
#13 U-tube 0.2 26 e e Seepage 8.7

temperature sensors were calibrated before the experiment [34].

The thermocouples were calibrated using a constant-temperature
water bath, ensuring an error of less than ±0.1
C. The un-
certainties dxi and relative uncertainties dRxi of the measured pa-
rameters can be obtained from Eqs. (4) and (5), where A is the
upper limit of the measuring range and gi is the accuracy grade. If a
parameter F is a function of a series of measured independent
variables xi, the relative uncertainty dRF for F can be obtained from
Eq. (6). Table 5 shows the relative uncertainties at a certain time
point when the water bath temperature is 36
C (#5) and the
vacuum degree is 1 kPa (#6) under stable working conditions. The
reason for the large error is that the small scale of the indoor test
leads to a small temperature difference between Tcooling,out and
Tcooling,in.

dxi ¼ A,gi (4)

d xi
dRxi ¼ (5)
xi
Fig. 4. Heat transfer rate of different materials.
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
 2
Pn vFi
i¼1 vxi dxi addition, it can be seen that q of the M-PE TPCT with time is more
dRF ¼ (6) stable than that of the PE TPCT. The residual ε is calculated between
F
qave and q, and then, the standard deviation s is calculated ac-
cording to ε. s is used to evaluate measures of dispersion of qave to q,
which can explain why the heat transfer rate of the M-PE TPCT is
3. Results and discussion more stable.
Fig. 5 shows the outer wall temperature of M-PE TPCT(a) and PE
3.1. Dependence of heat transfer rate on different TPCT construction TPCT(b) with time when the TPCT is running. The temperature
materials sensor intended to measure T120cm, which is 120 cm from the
bottom of the TPCT, failed during the experiments, and thus, no
Fig. 4 shows the heat transfer rate (q) of the TPCT for different data is shown. It can be seen that the heat absorption and heat
materials over time. As shown in Fig. 4, when the PE TPCT and M-PE release processes of the cycle destroy the isothermal property of
TPCT were under the same conditions, q gradually became stable the TPCT. During the running of the two kinds of TPCTs, the outer
after 360 s. During the stable stage, the average heat transfer rate wall temperature gradually decreases from the evaporation section
(qave) of the PE TPCT is 85.02 W. The internal HTCs in the evapo- to the condensation section. The temperature of the outer wall of
ration and condensation sections are 231.21 W/(m2$
C) and the two kinds of TPCT in the evaporation section is lower than the
250.98 W/(m2$
C), respectively. The qave of the M-PE TPCT is Tconst,in of 31
C, which indicates that the evaporation section ab-
183.47 W, and the internal HTCs in the evaporation and conden- sorbs heat well. In the evaporation section, T40cm is significantly
sation sections are 460.91 W/(m2$
C) and 471.77 W/(m2$
C), lower than T0cm and T20cm. This situation is more obvious for the M-
respectively. The thermal conductivity of the PE pipe is 0.33 W/ PE TPCT owing to the smaller thermal resistance of the pipe. In this
(m$
C), which is much lower than the 51.6 W/(m$
C) thermal experiment, the V is 12.5%, the liquid level of the working fluid is
conductivity of the galvanized steel pipe. Although the wall thick- approximately 25 cm. This is because the reflux of the liquid at
ness of the PE pipe is reduced to 1.5 mm, the qave and HTCs of the PE point T40cm will absorb a substantial amount of heat and decrease
TPCT are significantly lower than those of the M-PE TPCT. In the wall temperature of the TPCT. In addition, the temperature of

Table 5
Relative uncertainty of some working conditions.

Parameters Type of data Unit Water bath temperature is 36
C (#5) The vacuum degree is 1 kPa (#6)

Typical value Relative uncertainty Typical value Relative uncertainty

Inlet cooling water temperature (Tcooling,in) Measured C 10.11 0.99% 10.01 1.00%


Outlet cooling water temperature (Tcooling,out) Measured C 11.58 0.86% 11.02 0.91%
The flow rate of cooling water (Q) Measured m3/h 0.200 1.35% 0.207 1.30%
Heat transfer rate (q) Calculated W 343.00 9.71% 243.92 14.06%

402
W. Song, C. Zheng and J. Yang Renewable Energy 177 (2021) 397e410

Fig. 6. Heat transfer characteristics under different cooling water flow rates (Q).

Fig. 5. Outer wall temperature when the TPCT is running.

outer wall temperature of the condensation section under different
flow rates of cooling water (Q). It can be seen that without any
the outer wall of the condensation section is higher than the change in the properties of the M-PE TPCT itself when Q is
temperature of the cooling water at 10
C, indicating that the inner decreased, the qave decreases and the outer wall temperature of the
portion of the condensing section is continuously releasing heat. condensation section is increased. Q is the only independent vari-
T200cm in the M-PE TPCT is higher than T160cm and T180cm. As shown able. This indicates that the convective heat transfer rate capacity
in Fig. 2, because the cooling water outlet is lower than point T200cm, between the cooling water and the outer wall is enhanced when Q
the cooling water at the point T200cm cannot flow after absorbing a is increased. At the same time, the condensation rate of the working
large amount of heat; thus, the temperature of the wall continues to fluid is accelerated, which increases the q of the M-PE TPCT. The
increase here. Although this problem is also reflected in the PE radial heat transfer rate increases, resulting in a decrease in the
TPCT, it is difficult for the cooling water at point T200cm to be heated temperature of the outer wall.
up owing to less heat exchange in the borehole. According to Figs. 4 Fig. 6b shows the qave and its increase at different Q values. It can
and 5, it can be found that the variance of the M-PE TPCT is smaller be seen that with the same increase in Q, the increase in qave is
than that of the PE TPCT. In terms of heat transfer rate and stability, different. The processes of significantly increased qave range from
the M-PE TPCT is significantly better than the PE TPCT. For this 0.10 m3/h to 0.15 m3/h and 0.15 m3/h to 0.20 m3/h, and the qave
reason, the M-PE TPCT was selected in subsequent experiments. increases by 38.26 W and 34.24 W, respectively. By analyzing the
flow rate state of the cooling water in the water jacket, the Reynolds
number is 2332 when Q is 0.15 m3/h. The flow state in the water
3.2. Dependence of heat transfer rate on cooling water flow rate jacket is in the turbulent transition zone. When Q exceeds 0.20 m3/
h, as 0.20 m3/h to 0.25 m3/h and 0.25 m3/h to 0.30 m3/h, the
Fig. 6a shows the heat transfer rate (q) of the M-PE TPCT and the
403
W. Song, C. Zheng and J. Yang Renewable Energy 177 (2021) 397e410

increase in the amount of qave will be almost the same for every in p is 1.00 kPa from 1 kPa to 2 kPa and 2 kPa to 3 kPa, the decrease
increase of 0.05 m3/h, that is, 23.48 W and 20.76 W, respectively. in qave is different. The maximum reduction in the heat transfer rate
is as high as 83.52%. Fig. 10 shows the q of the M-PE TPCT under
3.3. Dependence of heat transfer rate on water bath temperature different filling ratios of the working fluid (V). It can be seen that the
qave during the stable stage when V are 12.5%, 25.0%, and 37.5% are
The heat transfer rates (q) of the M-PE TPCT for different tem- 183.47 W, 145.81 W, and 67.21 W, respectively. Similar to the results
peratures of the water bath (Tconst,in) with time are shown in Fig. 7. for different vacuum degree conditions, the increase in V is 12.5%
In the stable phase, Tconst,in is 26
C, 31
C, and 36
C, the average from 12.5% to 25.0% and 25.0% to 37.5%, the decrease in qave is
heat transfer rate (qave) is 157.96 W, 183.47 W, and 355.37 W, different. The maximum reduction in the heat transfer rate is as
respectively. When Tconst,in changes from 26
C to 31
C, qave in- high as 53.91%. This indicates that the change in p or V will affect
creases by 25.51 W. However, qave increases by 171.90 W when the performance of the M-PE TPCT when the working fluid is water.
Tconst,in changes from 31
C to 36
C. This indicates that the change Moreover, when p or V is higher, the performance of the M-PE TPCT
in Tconst,in will affect the performance of the M-PE TPCT when the is worse. In existing research, the recommended filling ratio is
working fluid is water. When Tconst,in is high, the performance of the approximately 15e30% for a conventional TPCT. However, in this
M-PE TPCT is high. experiment, a deeper liquid pool will cause a greater pressure at the
The T20cm values under different Tconst,in with time are shown in bottom. This will increase the saturation temperature of the
Fig. 8a. In the stable phase, the actual temperatures of the water working fluid at the bottom of the liquid pool. When the saturation
baths are 26.19
C, 31.19
C, and 36.19
C. At this time, the average temperature of the working fluid exceeds the soil temperature, the
temperatures of T20cm are 25.82
C, 30.81
C, and 35.60
C, working fluid cannot boil.
respectively. It can be seen that the average outer wall temperature The outer wall temperatures of the condensation section under
is lower than the actual temperature of the water bath when the M- different p and V with time when the M-PE TPCT is not running are
PE TPCT is running. This phenomenon indicates that the evapora- shown in Fig. 11a and Fig. 11b, respectively. It can be seen that when
tion section of the M-PE TPCT continuously absorbs heat. The the cooling water supply is stopped, the temperature of the outer
temperature difference can indicate the strength of heat absorp- wall of the condensation section can rise to a higher temperature
tion. Fig. 8b shows the relationship between qave and delta-T under faster. The temperature of the outer wall will be higher and far
different conditions, where delta-T is the temperature difference exceeds the saturated steam temperature corresponding to the
between the average Tconst,in and T20cm. When Tconst, in is 26
C, 31
C, previously set p after stopping operation when the qave is greater.
and 36
C, delta-T values are 0.37
C, 0.38
C, and 0.59
C, respec- Judging from the existing data, the saturation temperatures of
tively. It can be seen that the increasing trend of delta-T (repre- water are approximately 6.95
C, 17.54,
C and 24.11
C, respectively,
sented by the red solid line) and the increasing trend of the heat when p values are 1.00 kPa, 2.00 kPa, and 3.00 kPa, respectively.
transfer rate (represented by the red dashed line) are almost When p is 1 kPa, the temperature of the condensation section is
identical. This is the reason why even with the same water bath even close to 26
C about 16 min after the test is stopped. This
temperature difference of 5
C, the increase in heat transfer rate is shows that the steam in the M-PE TPCT is superheated, and it shows
different. that the pressure in the M-PE TPCT far exceeds the initial condition
of 1.00 kPa. This indirectly proves that the liquid pool boils
3.4. Dependence of heat transfer rate on vacuum degree and filling violently, which leads to a greater heat transfer rate. On the con-
ratio trary, when p is 3 kPa, although the p in the M-PE TPCT is set at
3.00 kPa, the saturated steam temperature should be close to
Fig. 9 shows the heat transfer rate (q) of the M-PE TPCT over time 24.11
C, and the liquid pool does not boil violently. This will result
under vacuum degrees (p) of 1 kPa, 2 kPa, and 3 kPa. It can be seen in too little steam, which will reduce q. From the experimental
that the average heat transfer rate (qave) during the stable stage is results, overheating is a good phenomenon that will improve the
233.94 W, 183.47 W, and 30.23 W, respectively, when p values are performance of the M-PE TPCT.
1 kPa, 2 kPa, and 3 kPa. It is worth noting that although the increase
3.5. Dependence of heat transfer rate on seepage or non-seepage
conditions

Fig. 12a shows the heat transfer rate (q) of the U-tube and M-PE
TPCT in the sandbox under non-seepage conditions with time. It
can be seen that q gradually became stable after 1000 s. In the
steady state, the average heat transfer rates (qave) of the U-tube and
M-PE TPCT are 41.80 W and 34.89 W, respectively. Because the
depth of the down tube is 1 m, the heat exchange rates per meter
are, respectively, 41.80 W/m and 34.89 W/m. In this experiment,
the qave of the U-tube is slightly higher than that of the M-PE TPCT
at approximately 19.8%. In fact, the length of the evaporation sec-
tion of the M-PE TPCT is 0.5 m. Therefore, the effective heat transfer
length of the M-PE TPCT is only 0.5 m, while the effective heat
transfer rate length of the U-tube is 1 m. The heat transfer area of
the U-tube is four times that of the M-PE TPCT. Fig. 12b shows the q
values of the U-tube and M-PE TPCT in the sandbox under seepage
conditions with time. If there is groundwater seepage in the soil,
the heat transfer capacity of the ground heat exchanger will be
greatly improved [35]. In a steady state, the qave values of the U-
tube and M-PE TPCT are 136.76 W and 111.75 W, respectively, the
Fig. 7. Heat transfer rate dependence on water bath temperature (Tconst,in). heat exchange rates per meter are 136.76 W/m and 111.75 W/m.
404
W. Song, C. Zheng and J. Yang Renewable Energy 177 (2021) 397e410

Fig. 8. Relationship of delta-T with heat transfer rate under different water bath temperatures (Tconst, in).

Fig. 9. Heat transfer rate under different vacuum degrees (p). Fig. 10. Heat transfer rate under different filling ratios (V).

Compared with Fig. 12a, it can be clearly observed that relying on Fig. 13a shows the initial temperature inside the sandbox at the
the flow of groundwater, the heat exchange of the two heat ex- beginning of the M-PE TPCT non-seepage experiment, and Fig. 13b
changers has been significantly improved. Among them, the qave of shows the temperature inside the sandbox after the experiment. It
the U-tube and M-PE TPCT increased by 327.18% and 337.49%, can be clearly seen that the sandbox has a good temperature
respectively. It is worth noting that regardless of whether the gradient in the initial stage of the experiment, as shown in Fig. 13b.
sandbox conditions involve seepage or not, the heat transfer rate of After the M-PE TPCT has run for 4500 s, around the M-PE TPCT
the M-PE TPCT is about 80% of that of the U-tube. In addition, the q evaporation section (30 < x < 50 cm, 0 < z < 50 cm), the soil
of the M-PE TPCT always oscillates regardless of seepage conditions. temperature dropped significantly. The area with the largest tem-
The reason for this is the low soil heat flux density or high filling perature drop is near point x ¼ 40 cm and z ¼ 25 cm, and the
ratio volume. Therefore, the production process of this type of TPCT temperature drop is approximately 2.21
C. Because the filling ratio
is only a preliminary exploration, and there is still room for of the working fluid in the M-PE TPCT is 12.5% and the liquid level of
improvement in terms of structure. the working fluid is 25 cm, M-PE TPCT has the strongest heat

405
W. Song, C. Zheng and J. Yang Renewable Energy 177 (2021) 397e410

Fig. 11. Outer wall temperature of condensation section.

Fig. 12. Heat transfers rate under sandbox experiments.

absorption capacity at 25e50 cm from the bottom, and the largest
temperature change. On the contrary, because the M-PE TPCT uses
the PE pipe as the pipe wall in the adiabatic section, the tempera-
which shows that the surrounding environmental temperature
ture change at the top of the sandbox is small. Fig. 14a shows the
hardly affects the soil.
initial temperature inside the sandbox at the beginning of the M-PE
The lower the soil temperature, the weaker the heat exchange
TPCT seepage experiment, and Fig. 14b shows the temperature in-
capacity of the ground heat exchanger. The presence of ground-
side the sandbox after the experiment. Because the groundwater
water has a great influence on the heat transfer performance of the
seepage environment is simulated, the overall temperature of the
ground heat exchanger and the surrounding soil, which is benefi-
sandbox tends to be the same at the beginning of the experiment,
cial to the long-term efficient operation of the heat exchanger
and the temperature is similar to the temperature at the bottom of
[36,37]. Comparing Figs. 12, 13b and 14b, it can be found that when
the sandbox under non-seepage conditions. The seepage water in
groundwater seepage occurs in the heat exchange conditions of the
the area of 50 < x < 70 cm and 0 < z < 10 cm has a relatively low
M-PE TPCT, it can effectively alleviate the decreasing trend of soil
temperature owing to the low seepage velocity. It can be seen that
temperature around the evaporation section of the M-PE-TPCT, and
the temperature-decrease area is in the evaporation section of the
the heat exchange capacity has also been significantly improved.
M-PE TPCT, and the temperature decrease is most evident near
This provides good heat exchange conditions for the long-term
point x ¼ 40 cm and z ¼ 25 cm; the decrease is approximately
stable operation of the M-PE-TPCT.
1.39
C in Fig. 14. Comparing the soil temperature before and after
the experiment in Figs. 13 and 14, it can be found that the heat
transfer of the M-PE TPCT does not break through the boundary of 4. Conclusions
the sandbox main body, and the boundary temperature of the
sandbox main body is consistent with the initial temperature, The aim of this study is to provide the heat transfer rate char-
acteristics of the proposed TPCT, consisting of a metal pipe and PE
406
W. Song, C. Zheng and J. Yang Renewable Energy 177 (2021) 397e410

Fig. 13. Temperature inside sandbox in non-seepage experiment under M-PE TPCT.

407
W. Song, C. Zheng and J. Yang Renewable Energy 177 (2021) 397e410

Fig. 14. Temperature inside sandbox in seepage experiment under M-PE TPCT.

408
W. Song, C. Zheng and J. Yang Renewable Energy 177 (2021) 397e410

pipe (M-PE TPCT), with water as the working fluid. A constant- financial interests or personal relationships that could have
temperature water bath test bench and a sandbox test bench appeared to influence the work reported in this paper.
were used to investigate the heat transfer characteristics with
respect to the pipe materials, cooling water flow rate, water bath Acknowledgments
temperature, vacuum degree, and filling ratio of the working fluid.
Both seepage and non-seepage conditions were investigated using This work was supported by the National Natural Science
the sandbox and compared with a traditional U-tube borehole heat Foundation of China (41602278), China Postdoctoral Science
exchanger. Several conclusions were drawn from this study, as Foundation (2016M601129), and North China University of Tech-
listed next. nology Programs (18XN154-006, 110052972027-031).

(a) The materials of the evaporation and condensation sections Nomenclature

will affect the heat transfer rate of the TPCT. Under the same
working conditions, when the evaporation section and
condensation section of the TPCT are made by PE pipes with a Symbols
wall thickness of 1.5 mm, the average heat transfer rate is q heat transfer of the two-phase closed thermosiphon
85.02 W. When the materials of the evaporation and heat exchange, W;
condensation sections are changed to galvanized steel pipes, qave average heat transfer of the two-phase closed
the heat transfer rate is 183.47 W. This shows that the ma- thermosiphon heat exchange, W;
terials in the evaporation section and the condensation sec- Q flow rate of cooling water, m3/h;
tion will significantly affect the heat transfer rate of the TPCT. Qave average flow rate of cooling water, m3/h;
(b) When the M-PE TPCT is running stably, the greater the flow p vacuum degree of two-phase closed thermosyphons
rate of cooling water, the greater the heat transfer rate. heat exchange, kPa;
However, when the Reynolds number of the cooling water in V filling ratio of the two-phase closed thermosyphons
the condensation section is in the turbulent transition zone, heat exchange working fluid, %;
the increase in heat transfer rate is the largest. In the laminar hc internal HTC in the condensation section, W/(m2$
C);
and turbulent flow zones, the cooling water flow rate in- he internal HTC in the evaporation section, W/(m2$
C);
creases, and the increase in heat transfer rate is small. de inner diameter of evaporation section, m;
(c) The heat transfer rate of the M-PE TPCT is affected by the dc inner diameter of condensation section, m;
temperature of the water bath, the vacuum degree, and the le length of evaporation section, m;
filling ratio of the working fluid. The reason for the change in lc length of condensation section, m;
the heat transfer rate can be analyzed through the change Tave,e average temperatures of the outer wall of the TPCT in
law of the outer wall temperature of the M-PE TPCT. The the evaporator section,
C;
experimental results show that when the cooling water flow Tave,a average temperatures of the outer wall of the TPCT in
rate is 0.2 m3/h, with a water bath temperature of 36
C, the adiabatic section,
C;
vacuum degree of 1.00 kPa, and filling ratio of 12.5%, the M- Tave,c average temperatures of the outer wall of the TPCT in
PE TPCT has the highest average heat transfer rate of the condensation section,
C;
355.37 W. T0cm, T20cm … T200cm outer wall temperature of TPCT, with 0 cm at
(d) The results of the sandbox experiment show that the average the bottom of the borehole,
C;
heat transfer rate of the M-PE TPCT is 34.89 W/m and Tcooling,in inlet water temperature of cooling water in the jacket,
111.75 W/m under the conditions of no seepage and seepage,
C;
respectively. The seepage conditions greatly increase the Tcooling,out outlet water temperature of cooling water in the jacket,
heat transfer rate of the M-PE TPCT by 337.49%. When the
C;
heat transfer length of the M-PE TPCT is half that of the U- Tconst,in temperature of the water bath,
C;
tube, the heat transfer rate of the M-PE TPCT is approxi- delta-T temperature difference between average Tconst,in, and
mately 80% of that of the U-tube. Under seepage conditions, T20cm,
C.
the difference in the heat transfer rate remains the same. The
experiment showed that it is feasible to use the new M-PE Acronyms
TPCT as an underground heat exchanger. TPCT Two-phase closed thermosyphon heat exchanger;
PE TPCT Polyethylene TPCT;
The M-PE TPCT uses a new structure, which is composed of a M-PE TPCT Metal-polyethylene TPCT;
galvanized steel pipe and PE pipe, and uses water as the working GSHP Ground source heat pump;
fluid. The TPCT with this structure can reduce the cost, the working FECSCW Forced external circulation standing column well;
fluid is environmentally friendly, safe, and convenient for trans- SCW Standing column well;
portation, and the production difficulty is low. Because the new PRW Pumping and recharging well;
structure is in the indoor test stage, a shorter design length leads to COP Coefficient of performance;
larger errors, but the errors will be further reduced in actual en- PE Polyethylene;
gineering applications. Although the change in design length has a DC Direct current.
great influence on the heat transfer performance of the TPCT, this
work can be used as a basis for further research on M-PE TPCTs. We
References
will also explore the effect of M-PE TPCTs in practical engineering
applications. [1] H. Zoelly, Heizfahren, Patent Office, Switzerland, 1912, p. 59350.
[2] R. DiPippo, Geothermal Energy as a Source of Electricity, 1980. Washington
Declaration of competing interest D.C.
[3] Anonymous Fireless, Furnace dit Pumps Heat from Earth to House, Life
Magazine, 1948.
The authors declare that they have no known competing [4] Y.D. Li, X. Zhang, Review on the research and application of ground source

409
W. Song, C. Zheng and J. Yang Renewable Energy 177 (2021) 397e410

heat pump in China and abroad, Power Generation Technology 85 (23) (2020) geothermal probes according to the principle of a heat pipe, KI-Kalte Luft
4e7. Klimatechnik 40 (2004) 54e61.
[5] N. Molina-Giraldo, P. Blum, K. Zhu, et al., A moving finite line source model to [24] H. Kruse, H. Russmann, Novel CO2-heat pipe as earth probe for heat pumps
simulate borehole heat exchangers with groundwater advection, Int. J. Therm. without auxiliary pumping energy, in: Proc. 8th International IEA Heat Pump
Sci. 50 (2011) 2506e2513. Conference, 2005. Las Vegas.
[6] D.A. Ball, R.D. Fischer, D. Hodgett, Design methods for ground-source heat [25] H. Kruse, H. Russmann, The status of development and research on CO2 earth
pumps, Build. Eng. 89 (1983) 416e440. heat pipes for geothermal heat pumps, International High Performance
[7] J.E. Bose, C.W. Ledbetter, J.R. Partin, Experimental results of a low-cost solar- Buildings Conference (2010) 51.
assisted heat pump system using earth coil and geo-thermal well storage, in: [26] J. Acuna, B. Palm, R. Khodabandeh, et al., Distributed temperature measure-
The 4th Annual Heat Pump Technology Conference, 1979. Stillwater, OK, USA. ments on a U-pipe thermosyphon borehole heat exchanger with CO2, in: 9th
[8] J. Oliver, H. Braud, Thermal exchange to earth with concentric well pipes, IIR Gustav Lorentzen Conference, 2010. Sydney.
Transactions of the ASAE American Society of Agricultural Engineers 24 (4) [27] J.C. Ebeling, X. Luo, S. Kabelac, et al., Dynamic simulation and experimental
(1981), 0906-0910. validation of a two-phase closed thermosyphon for geothermal application,
[9] H. Braud, H. Klimkowski, J. Olivier, Earth source heat exchanger for heat Propulsion and Power Research 6 (2) (2017) 107e116.
pumps, ASAE Transactions (26) (1983) 1818e1822. [28] J.C. Ebeling, S. Kabelac, S. Luckmann, et al., Simulation and experimental
[10] C. Tan, E. Kush, Semi-closed loop standing column well groundwater source validation of a 400 m vertical CO2 heat pipe for geothermal application, Heat
heat pump[EB/OL], GHPC website, www.geoexchange.org. Mass Tran. 7 (2014).
[11] C.D. Orio, Geothermal heat pumps and standing column wells, Trans. Geoth. [29] H. Lim, C. Kim, Y. Cho, et al., Energy saving potentials from the application of
Resour. Counc. (18) (1994) 375e379. heat pipes on geothermal heat pump system, Appl. Therm. Eng. 126 (2017)
[12] C.D. Orio, Design, use & example of standing column wells, in: IGSPHA 1191e1198.
Technical Meeting, 1995. Stillwater, OK, USA. [30] X.Y. Wang, H. Liu, Y.F. Wang, et al., CFD simulation of dynamic heat transfer
[13] C.D. Orio, Geothermal heat pump applications industrial/commercial, Energy rate behaviors in super-long thermosyphons for shallow geothermal appli-
Eng. J. Assoc. Energy Eng. 96 (3) (1999) 58e66. cation, Appl. Therm. Eng. 174 (2020) 115295.
[14] C.D. Orio, C.N. Johnson, K.D. Poor, Geothermal standing column wells: ten [31] R. Rieberer, K. Mittermayr, H. Halozan, CO2 Thermosyphons as heat source
years in a new england school, Build. Eng. 112 (2) (2006) 57e64. system for heat pumps-4 years of experience, in: 8th IEA Heat Pump Con-
[15] C.D. Orio, C.N. Johnson, S.J. Rees, et al., A survey of standing column well in- ference, 2005.
stallations in North America, Build. Eng. 111 (2) (2005) 109e121. [32] L.M. Ma, L.L. Shang, D. Zhong, et al., Experimental investigation of a two-phase
[16] S.N. Sorensen, J. Reffstrup, Prediction of long-term operational conditions for closed thermosyphon charged with hydrocarbon and Freon refrigerants, Appl.
single-well groundwater heat pump plants[C], in: The 27th Intersociety En- Energy 207 (2017) 665e673.
ergy Conversion Engineering Conference, 1992. San Diego, CA, USA. [33] W. Song, L. Ni, Y. Yao, Analyses of different operation modes of pumping and
[17] S. Xu, L. Rybach, Utilization of shallow resources performance of direct use recharging well using an indoor sandbox test, Energy Build. 208 (1) (2020)
system in beijing, Trans. Geoth. Resour. Counc. 27 (2003) 115e118. 109660.
[18] K. Ochsner, Carbon dioxide heat pipe in conjunction with a ground source [34] C. Shen, Y. Wang, Z.L. Zhao, et al., Decoupling analysis on the variations of
heat pump (GSHP), Appl. Therm. Eng. 28 (2008) 2077e2082. liquid velocity and heat flux in the test of fouling thermal resistance, Int. J.
[19] F. Hartmann, R. Behrend, A. Hantsch, et al., Numerical investigation of the Heat Mass Tran. 123 (2018) 227e238.
performance of a partially wetted geothermal thermosyphon at various po- [35] S. Yan, S.F. Li, L.H. Dai, Study of characteristics of soil temperature variation
wer demand schemes, Geothermics 55 (2015) 99e107. and recovery under intermittent operation of ground-source heat pump,
[20] R. Rieberer, Naturally circulating probes and collectors for ground-coupled J. Dalian Univ. Technolgy 52 (2012) 350e356, 03.
heat pumps, Int. J. Refrig. 28 (2005) 1308e1315. [36] R. Fan, Y.Q. Jiang, Y. Yao, et al., A study on the performance of a geothermal
[21] R. Rieberer, K. Mittermayr, H. Halozan, CO2 thermosyphons as heat source heat exchanger under coupled heat conduction and groundwater advection,
system for heat pumps d 4 years of market experience, in: Proceedings of the Energy 32 (11) (2007).
8th IEA Heat Pump Conference, 2005. Las Vegas, NV, USA. [37] R. Fan, Y.Q. Jiang, Y. Yao, et al., Theoretical study on the performance of
[22] R. Rieberer, K. Mittermayr, H. Halozan, CO2 Heat Pipe for Heat Pumps, IIR/IIF- anintegrated ground-source heat pump system in a whole year, Energy 33
Commission B1, B2, E1 and E2, Guangzhou, China, 2002, pp. 200e207. (11) (2008).
[23] H. Kruse, H. Russmann, C. Stadtlander, Development of CO2-driven

410