Applied Energy 88 (2011) 4113–4119

Contents lists available at ScienceDirect

Applied Energy
journal homepage: www.elsevier.com/locate/apenergy

Heat transfer analysis of pile geothermal heat exchangers with spiral coils
Ping Cui a,b,c,⇑, Xin Li c, Yi Man c, Zhaohong Fang b
a
Key Laboratory of Renewable Energy Utilization Technologies in Buildings, Ministry of Education, Jinan, China
b
School of Thermal Energy Engineering, Shandong Jianzhu University, Jinan, China
c
Shandong Key Laboratory of Building Energy-saving Technique, Jinan, China

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

Article history: This paper investigates the transient heat conduction around the buried spiral coils which could be
Received 26 September 2010 applied in the ground-coupled heat pump systems with the pile foundation as a geothermal heat exchan-
Received in revised form 22 February 2011 ger. A transient ring-coil heat source model is developed, and the explicit analytical solutions for the tem-
Accepted 29 March 2011
perature response are derived by means of the Green’s function theory and the image method. The
Available online 20 April 2011
influences of the coil pitch and locations are evaluated and discussed according to the solutions. In addi-
tion, comparisons between the ring-coil and cylindrical source models give that the improved finite ring-
Keywords:
coil source model can accurately describe the heat transfer process of the pile geothermal heat exchanger
Ground-coupled heat pump
Pile geothermal heat exchanger
(PGHE). The analytical solutions may provide a desirable and better tool for the PGHE simulation/design.
Heat transfer Ó 2011 Elsevier Ltd. All rights reserved.
Spiral coil
Ring-coil source

1. Introduction engineers in the GCHP industry. This new system is known as

the ‘‘Pile Geothermal Heat Exchanger (PGHE) system’’. The piles
The ground-coupled heat pump (GCHP) technology is one of the can not only support the building structure but also act as heat
most energy efficient ways to provide both space heating and cool- exchangers. Each pile can carry a closed-loop pipe connected to a
ing for buildings because the underground environment experi- heat pump on the ground for heating, cooling and hot water
ences less temperature fluctuation than the ambient air supply. Pipes may be buried in concrete piles in configurations of
temperature. In a GCHP system, heat is extracted from or rejected U-tubes or spiral coils. The pile grout material, mainly concrete,
to the ground via a geothermal heat exchanger (GHE) through provides good tight contact between the buried pipes and the piles
which pure water or anti-freezing solution circulates. The com- and between the piles and the surrounding soil as well. Therefore it
monly used GHE in closed loop systems typically consists of can significantly reduce the thermal contact resistance, and im-
high-density polyethylene (HDPE) U-tubes installed in vertical prove the heat transfer efficiency of the PGHE. The most competi-
boreholes which are called vertical GCHP systems. A number of tive advantage of this system is the considerable reduction of the
vertical GCHP systems have been installed around the world initial cost and the plot of land for the borehole field.
including in China due to the increasing awareness of environmen- A few studies on PGHE systems have been recently reported in
tal issues and growing energy crisis [1,2]. However, the main literature, mostly by either experiments or numerical simulations.
restriction against the GCHP application in urban areas is that Morino and Oka (1994) first employed steel piles, which were used
the system requires a large plot of land for installation of GHEs, for building foundations, as heat exchangers [3]. In their study, the
and this is usually difficult in cities. Besides, larger GHEs signifi- heat transfer characteristics of the steel pile equipped with two
cantly increase the initial cost of the GHE installation. Therefore, vertical tubes were evaluated by means of experiments and a
the high installation cost of excessively large GHEs and the limited three-dimensional finite difference method. Pahud (1996) devel-
urban land area have restricted to a large extent the wide applica- oped a simulation tool (called PileSIM) based on a TRNSYS environ-
tions of this technology in densely-populated areas. ment for heating and cooling, which used heat exchanger piles
Recently, a novel GHE system which utilizes the foundation with vertical U-tube pipes [4]. The simulation tool employed the
piles of a building as the means of transferring heat to and from Duct Ground Heat Storage Model (DST) which was solved by
the ground has attracted great interest among researchers and means of numerical methods. Later, a heat exchanger pile system
was designed by using the simulation tool for heating and cooling
⇑ Corresponding author at: Key Laboratory of Renewable Energy Utilization a terminal building at Zurich Airport [5]. Laloui et al. (2006) stud-
Technologies in Buildings, Ministry of Education, Jinan, China. ied the behaviors of a heat exchanger pile for thermo-mechanical
E-mail address: sdcuiping@gmail.com (P. Cui). loads, especially increased loads on the pile caused by thermal ef-

0306-2619/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.apenergy.2011.03.045
4114 P. Cui et al. / Applied Energy 88 (2011) 4113–4119

Nomenclature

a thermal diffusivity (m2 s1) Z dimensionless coordinate

b coil or spiral pitch (m)
B dimensionless coil pitch Greek symbols
c specific heat (J kg1 K1) h temperature excess (°C)
Fo Fourier number H dimensionless temperature excess
h1, h2 pile depth (m) q density (kg m3)
H dimensionless depth s time (s)
k thermal conductivity (W m1 K1) u angular coordinate (rad)
q heating rate (W)
ql heating rate per length (W m1) Subscripts
R dimensionless radius c cylindrical source
r radial coordinate (m) f finite length
r0 pile radius (m) i infinite length
t temperature (°C) r ring-coil source
t0 initial temperature (°C) s single ring-coil source
z axial coordinate (m)

fects, using a coupled multi-physical finite element model [6]. the models for the conventional boreholes are commonly used to
Yasuhiro et al. (2007) investigated the field performance of a PGHE give a rough estimation in the PGHE applications. However, the
system using friction piles as heat exchangers for space heating heat transfer characteristics of the PGHE system are quite different
which was built in Japan in 2000 [7]. Sekine et al. (2007) developed from those of conventional vertical boreholes. Piles are much
a GCHP system that utilized the cast-in-situ concrete pile founda- thicker in diameter but shorter in depth than boreholes. The heat
tion embedded in U-tubes as heat exchangers [8]. transfer coils are inside the pile concrete but disposed within about
In the previous studies, the heat exchanger pipes in practice in 10 cm of the pile’s outer surface. Obviously, either the line source
the PGHE systems are U-tube pipes, including single, double, triple models [9–12] or the ‘‘hollow’’ cylindrical model [13] is no longer
or even quadruple U-tubes. Few studies on the PGHE with spiral valid in this case due to their respective simplifying assumptions.
coils have been carried out due to the complexity of its heat trans- Based on the classical models, our research group has recently pro-
fer process. However, the spiral coil configuration has the advanta- posed a new model, i.e. the ‘‘solid’’ cylindrical source model [14],
ges of more heat transfer area in a certain pile and better flow and takes the PGHE characteristics into proper consideration, as
pattern without air chocking in the pipes compared with the serial shown in Fig. 2. In this model, the cylinder is filled with the med-
or parallel U-tubes in the pile. In addition, the spiral coil system ium identical to that out of the cylinder rather than a cavity, and
can reduce the complexity of the pipe connections and decrease therefore the whole domain is composed of a homogeneous med-
to a certain extent the thermal ‘‘short-circuit’’ between supply ium. A heat source shaped in a cylindrical surface of a radius r0 is
and return pipes. The schematic diagram of a GCHP system with buried in the medium with its axis being coincident with z-axis.
spiral coils buried in a pile foundation is illustrated in Fig. 1. The thickness, mass and heat capacity of the heat source are ne-
Due to its short application history and geometrical characteris- glected. The heating rate per length of the cylindrical source, ql,
tics, few analytical models for the PGHE have been developed and is constant since a starting instant, s = s0 . The 2-D model for the
smaller depth-to-diameter ratio of the pile was also developed to
investigate the axial heat flow in our previous work.
As discussed in our previous paper [14], the temperature rise
caused by an infinite ‘‘solid’’ cylindrical heat source at any location
of the coordinate (r, z) and at the instant s can be obtained accord-
ing to the Green function theory as:

h1

h2
2r 0
z
Fig. 2. Schematic diagram of the PGHE (left: spiral coils; right: finite ‘‘solid’’
Fig. 1. Schematic diagram of a GCHP system with PGHEs. cylindrical source model).
 P. Cui et al. / Applied Energy 88 (2011) 4113–4119 4115

Z s    
ql 1 r 2 þ r 20 rr0 of a radius r0 with the pitch b being coincident with z-axis, as
hc;i ¼ exp   I0 ds0 ð1Þ
4pk 0 ss 0 4aðs  s0 Þ 2aðs  s0 Þ shown in Fig. 3. In order to analyze the thermal effect of the heat
Rp transfer along the axis on the total heat transfer efficiency, both
where I0 ðxÞ ¼ p1 0 expðx cos u0 Þdu0 is the modified Bessel function the infinite and finite ring-coil source models are discussed in this
of the zero order, hc;i ¼ t c;i  t0 is the temperature excess. paper.
In the 2-D model, the cylindrical heat source is considered with The analysis on the heat transfer around the buried ring-coil
a limited length, stretching from h1 to h2 in the Z direction. The pipes are based on the following assumptions.
image method was used to obtain the solution of this problem
[14], as below: (1) The ground is regarded as a homogeneous medium, both
Z s     inside and outside the cylindrical circumference; and its
q 1 rr 0 r 2 þ r 20
hc;f ¼ l I
0 0 2aðs  s0 Þ
exp  thermophysical properties do not change with temperature.
8pk 0 ðs  s Þ 4aðs  s0 Þ
( " # " # (2) The medium has a uniform initial temperature, t0.
z  h2 z  h1 (3) The ring-coils are assumed to be numerous ring-coil heat
 erfc pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi  erfc pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2 aðs  s0 Þ 2 aðs  s0 Þ sources. Each of them emits heat at the intensity of qlb since
" # " #)
a starting instant s = s0 . The thickness, mass and heat capac-
z þ h2 z þ h1
þ erfc pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi  erfc pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ds0 ð2Þ ity of the heat source are neglected.
2 aðs  s0 Þ 2 aðs  s0 Þ

While the ‘‘solid’’ cylindrical source model has made marked For the ring-coil source model of finite length additional
progress from the classical line source or ‘‘hollow’’ cylindrical assumptions are taken:
source models, it fails to distinguish the effect of the spiral pitches
because the coil is simplified as a continuous cylindrical surface. In (4) The ground is assumed to be a semi-infinite medium, and its
practice, the temperature distribution fluctuates significantly along boundary, i.e. the ground surface, keeps a constant temper-
the axial direction in the vicinity of the cylindrical surface due to ature t0 (same as its initial one) throughout the period
the non-integrity of the heat source. This feature is of great impor- concerned.
tance in analyzing the temperature rise of the buried pipe, and, (5) The ring coils, are buried perpendicular to the boundary,
then, the circulating fluid in it. As a result, new models are needed stretching from depth h1 to depth h2.
for better understanding and analyzing the heat transfer around a
buried coil. 2.1. Temperature response by a single ring-coil heat source
An advanced model is established in this paper, and solved ana-
lytically to better describe the heat transfer process of the PGHE To develop the analytical model of the spiral coils, studying a
with spiral coils. Thus, the main objective of this study is to provide single ring-coil heat source is the basic and simple starting point
a more accurate algorithm which can be further used to calculate and then introducing other complications step by step. The ring
the temperature of the fluid in the spiral coil pipes and design coil is supposed to be located on the plane z = z0 with its axis being
the PGHE systems. coincident with z-axis. It emits a continuous heating rate of q since
a starting instant, s = s0 .
Based on the governing equation of the transient heat conduc-
2. Ring-coil heat source model and solutions
tion along with the given boundary and initial conditions, the
two-dimensional heat conduction problem in the infinite medium
An advanced model, i.e. the ring-coil source model, is proposed
can be formulated in the cylindrical coordinate:
to take into account the discontinuity of the heat source and the
2 
impact of the coil pitches. The spiral coils buried in the pile are 2 0
@h
@s
¼ a @@r2h þ 1r @h
@r
þ @@z2h þ q dðrr 0 ;zz Þ
2pr0 qc
; for 0 < r < 1; 1 < z < 1; s > s0
simplified as a number of separated rings on the cylindrical surface
h ¼ 0; for 0 < r < 1; 1 < z < 1; s ¼ s0
@h
@r
¼ 0; for r ¼ 0; 1 < z < 1; s > s0
h ! 0; for r ! 1; 1 < z < 1; s > s0
h1 b/2 h ! 0; for z ! 1
b ð3Þ
Thus, the temperature response at the location (r, z) in the med-
ium to such a single ring source can be obtained according to the
Green’s function theory [15,16]:
Z s Z 2p
q
hr;s ¼ ds0 Gdu0
2pqc 0 0
Z s
q 1
¼
8qc 0 ½paðs  s0 Þ3=2
" #  
r 2 þ r20 þ ðz  z0 Þ2 rr 0
 exp  I 0 ds0 ð4Þ
4aðs  s0 Þ 2aðs  s0 Þ

where the Green’s function in the cylindrical coordinates is given

h2 as:
" #
2r 0 1 r 2 þ r 02  2rr0 cosðu  u0 Þ þ ðz  z0 Þ2
G¼  exp 
z 8½paðs  s0 Þ3=2 4aðs  s0 Þ

Fig. 3. The finite ring-coil source model. ð5Þ

4116 P. Cui et al. / Applied Energy 88 (2011) 4113–4119

Z s "
By introducing the following dimensionless variables: H ¼ khqr0 , Xm Z 2p
qb
0 0
Fo ¼ ar2s, Fo0 ¼ ars2 , R ¼ rr0 , Z ¼ rz0 , Z 0 ¼ rz0 , the dimensionless temperature hr;f ¼ l ds0
Gðz0 ¼ h1 þ nb þ 0:5bÞdu0
0 0
2pqc 0 n¼0 0
excess at (R, Z) can be expressed as a function of the following #
m Z 2p
X
dimensionless variables:  0
Gðz ¼ h1  nb  0:5bÞdu 0

Z Fo   n¼0 0
1 1 R Z s   
Hr;s ¼ I0 0 ql b 1 rr 0 r 2 þ r20
8p3=2 ðFo  Fo0 Þ3=2 2ðFo  Fo Þ
0 ¼ I
 exp 
3=2 0
" # " # 8qc 0 ½paðs  s0 Þ 2aðs  s Þ
0 4aðs  s Þ 0

R2 þ 1 ðZ  Z 0 Þ2 ( " #
 exp  exp  dFo
0
ð6Þ Xm
ðz  h1  nb  0:5bÞ 2
4ðFo  Fo0 Þ 4ðFo  Fo0 Þ  exp 
n¼1
4aðs  s0 Þ
" #)
ðz þ h1 þ nb þ 0:5bÞ2
2.2. Infinite ring-coil source model  exp  ds0 ð9Þ
4aðs  s0 Þ
The infinite ring-coil source model is firstly studied, which
means that the ring-coil source is assumed to be infinite in the Define H1 ¼ h1
r0
and H2 ¼ h2
r0
, then Eq. (9) can be rewritten as follows:
longitudinal direction. The z-coordinates of the ring coils are Z  
Fo
defined to be z0 = ± (n + 0.5)b, where n = 0, 1, 2, . . ., +1. As a conse- B 1 R
Hr;f ðR; Z; FoÞ ¼ I
3=2 0 0
quence, the overall temperature response at a random point in 8p3=2 ðFo  Fo0 Þ
0 2ðFo  Fo Þ
" #
the medium to all the ring sources can be determined as the
R2 þ 1
sum of all the individual temperature rises caused by each ring-  exp 
4ðFo  Fo0 Þ
coil source. ( " #
Xm
ðZ  H1  ðn þ 0:5ÞBÞ2
X Z s Z 2p  exp 
ql b 0
4ðFo  Fo0 Þ
hr;i ¼ ds0 Gðz0 ¼ nb  0:5bÞdu0 n¼0
" #)
2pqc n¼1 0 0
ðZ þ H1 þ ðn þ 0:5ÞBÞ2 0
!  exp  dFo ð10Þ
1 Z s
X Z 2p 4ðFo  Fo0 Þ
0 0 0
þ ds Gðz ¼ nb þ 0:5bÞdu
n¼0 0 0

1 Z s  
ql b X 1 rr 0
¼ I 0 3. Comparisons and discussion
8qc n¼0 0 ½paðs  s0 Þ3=2 2aðs  s0 Þ
( " # 3.1. Comparisons of infinite and finite ring-coil source models
r 2 þ r 20 þ ðz  nb  0:5bÞ2
 exp 
4aðs  s0 Þ
The temperature responses of the infinite and finite ring-coil
" #) source models can be calculated easily according to Eqs. (8) and
r 2 þ r 20 þ ðz þ nb þ 0:5bÞ2
þ exp  ds0 ð7Þ (10), respectively, and the computation results are illustrated in
4aðs  s0 Þ
Figs. 4 and 5.
Taking an example of infinite ring-coil source with B = 1, the
After introducing another dimensionless variable of B ¼ rb0 , Eq. temperature distributions at the cylindrical surfaces of R = 1,
(7) can be rewritten in the dimensionless format: R = 0.5 and R = 1.2 along the Z direction are described in Fig. 4. In
general, the temperature rises fluctuate considerably in the vicinity
1 Z
X Fo  
B 1 R to the ring-coil sources and shows a periodic variation in the Z
Hr;i ðR; Z; FoÞ ¼ I 0 0
8p3=2 n¼0 0 ðFo  Fo0 Þ3=2 2ðFo  Fo Þ direction, which is due to the discontinuity of the ring-coil sources
" #( " # in the longitudinal direction. The temperature at the cylindrical
R2 þ 1 ðZ  nB  0:5BÞ2
 exp  exp 
4ðFo  Fo0 Þ 4ðFo  Fo0 Þ
-1.5
" #) R=1
ðZ þ nB þ 0:5BÞ2 0 R=0.5
þ exp  dFo ð8Þ
4ðFo  Fo0 Þ -1 R=1.2
Dimensionless depth (Z)

-0.5
2.3. Finite ring-coil source model

The effects of heat flow through the top and bottom ends of the 0
pile are neglected in the above infinite model, and therefore it is
inadequate for the long-term operation of the PGHE especially in
0.5
view of the much smaller depth-to-radius ratio of the pile than that
of the borehole. While keeping the ring-coil source simplification,
the spiral coil is taken as a finite ring-coil source buried in a 1
semi-infinite medium, stretching from h1 to h2 from the boundary
of the ground surface. The coil is then approximated as m pieces of 1.5
rings, where m = int [(h2  h1)/b]. Again, the images of the ring 0.21 0.22 0.23 0.24 0.25 0.26 0.27 0.28 0.29
coils with negative heating rate qlb are set on symmetry to the Dimensionless temperature rise Θr,i
boundary in order to keep the constant temperature of the ground
surface, and the solution for the ring-coil source model of finite Fig. 4. Temperature distribution along the Z-axis with different R as parameter
length is expressed as: (B = 1, Fo = 10).
 P. Cui et al. / Applied Energy 88 (2011) 4113–4119 4117

0 perature at the point with R = 1 and Z = 0 is taken as the represen-

tative temperature of the coil surface. For the finite model, the
2
temperature at the middle of the pile depth (R = 1, Z = (H2 + H1)/
4 2) is computed for comparison purpose.
Dimensionless depth (Z)

6
The representative temperatures calculated from the two mod-
els are illustrated in Fig. 6. In general, the infinite model yields a
8 relatively higher temperature response compared with the finite
10 model because of its neglect of the heat transfer through the top
and bottom ends of the pile. The infinite ring-coil model results
12 in a good agreement with the finite model for a short initial period,
14 but remarkable discrepancy of the infinite model from the finite
R=1 Fo=0.1 one appears due to the assumption of the infinite heat source with
16 R=1 Fo=1 respect to long-term duration. Obviously, the shorter the pile is the
R=1 Fo=10
18 larger discrepancy between them occurs. When the time tends to
R=0.5,Fo=10
infinity, the temperature response of the infinite model also tends
20
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 to infinity as the finite model reaches a steady-state one which cor-
Dimensionless temperature rise Θr,f responds to the actual heat transfer mechanism. This feature dem-
onstrates the importance to take the length effect into account
Fig. 5. Temperature profiles along the pile depth with different Fo (with H1 = 2, regarding the long-term operation of the PGHEs. In the PGHE appli-
H2 = 12, B = 1). cations, the depth-to-diameter ratio of the heat source is usually
much smaller than that of the vertical boreholes, so the finite heat
source model should be employed in the thermal analysis of the
surface of ring coils with R = 1 appears an extraordinary fluctua- PGHEs.
tion, and tends to infinity at locations, where the ring sources exist.
For a PGHE with H1 = 2, H2 = 12, and B = 1, the temperature re-
3.2. Comparisons of infinite ring-coil source and ‘‘solid’’ cylindrical
sponses on the cylindrical surface of the ring-coil source (i.e. R = 1)
source models
along the pile depth at different instants (Fo) are computed and
plotted in Fig. 5. As shown in Fig. 5, for the case of Fo = 0.1, the tem-
The ring-coil source model becomes the ‘‘solid’’ cylindrical
perature distribution at R = 1 shows a quite similar periodic varia-
model when B = 0, where the temperature distribution is indepen-
tion to that of the infinite model (as seen in Fig. 4) except points
dent of Z coordinate for the case of infinite heat source. Fig. 7 com-
near the top and bottom ends of the pile. This indicates that the
pares the temperature distributions along the radial direction of
influence of the axial heat conduction caused by the boundary
infinite ring-coil source and the infinite ‘‘solid’’ cylindrical source
and finite length of the heat source is limited to the two ends of
(B = 0) models with different parameters of Z and B. It can be seen
the pile in relatively short periods of time. For long-term operation,
from Fig. 7 that the two models show a similar trend of the temper-
however, the impact of heat dissipation through the ends will pen-
ature variation along the radial direction. Each temperature profile
etrate deeper into the whole depth of the pile, and then the signif-
from the two models reaches an apex at the circumference of the
icant deviations from the temperature distribution of the infinite
heat source, i.e. R = 1, at any Z-level, and heat flows, then, to both
model occur, as shown in Fig. 5.
sides of the cylindrical circumference.
Even though the temperatures in both the infinite and finite
The temperature distribution of the ring-coil model fluctuates
ring-coil models vary with time and pile depth, as shown in Figs. 4
along the Z direction because of the discontinuous ring coils. As
and 5, the radical differences between them increase with the time
shown in Fig. 7, the temperature rise at the Z-level away from
elapses, especially for long-term periods. In view of the compari-
the ring coil may be lower than that of the cylindrical model, such
sons between the two models, the temperature at the mid-point
as Z = 0, whereas, the temperature rise near the ring-coil source
of two ring-coil sources on the cylindrical surface of R = 1 is chosen
exhibits a higher value than that of the cylindrical model. The
as a representative one. For the infinite ring-coil model, the tem-

0.16
0.8
Dimensionless temperature rise Θ

0.14
0.7
Dimensionless temperature rise Θ

Infinite ring-coil model 0.12

0.6

0.5 0.1
H=30
0.4 0.08

H=20 B=0 cylindrical source

0.3 0.06
B=0.4 Z=0
0.2
H=10 B=0.8 Z=0
0.04
B=0.4 Z=0.2
0.1 0.02 B=0.8 Z=0.4

0 0 0.5 1 1.5 2
-2 0 2 4
10 10 10 10
Dimensionless radius (R)
Dimensionless time (Fo)
Fig. 7. Temperature profiles along radial direction with different Z and B values
Fig. 6. Dimensionless temperatures vs. Fo from finite and infinite ring-coil models. (Fo = 1).
4118 P. Cui et al. / Applied Energy 88 (2011) 4113–4119

0.7 0
B=0 cylindrical source
Dimensionless temperature rise Θ

B=0.5 Z=0 2
0.6

Dimensionless depth (Z)

B=2 Z=0
4
B=0.5 Z=0.4*B
0.5
B=2 Z=0.4*B
6
0.4 RCS R=1
8
CS R=1
0.3 10 RCS R=0.5
CS R=0.5
0.2 12 RCS R=1.2
CS R=1.2
14
0.1 RCS R=1.5
16 CS R=1.5

0 0 0.05 0.1 0.15 0.2

-2 -1 0 1 2 3
10 10 10 10 10 10 Dimensionless temperature rise Θ
Dimensionless time (Fo)
Fig. 9. Comparisons of temperature profiles along pile depth between finite
Fig. 8. Comparisons of temperature responses on the cylindrical surface of R = 1 cylindrical and ring-coil source models (where RCS = ring-coil source and CS = cylin-
from infinite ring-coil and cylindrical sources with different pitches. drical source; H1 = 2, H2 = 12, B = 1, Fo = 1).

highest temperature rise occurs in the plane of Z = 0.4, where the the cylindrical model with large amplitude near the ring coil, and
ring-coil source is located, and the temperature obtained from the amplitude of the fluctuation gradually abates away from the
the ring-coil model at R = 1 tends to infinite due to the assumption cylindrical surface of the ring coil in the radial direction.
of the line source nature of the rings. According to the computation results and the comparisons, it is
Fig. 7 also depicts different temperature profiles with different obvious that the ring-coil model can better describe the heat con-
coil pitches in order to better investigate the effect of the ring-coil duction in the PGHEs with spiral coils compared with the ‘‘solid’’
density on the temperature distribution along the axis direction. It cylindrical model owing to its more realistic assumptions.
gives that the deviation of the cylindrical model from the ring-coil
model dramatically increases with the rise in the coil pitch for the
same heat intensity per length of the coils. For a smaller pitch, such 4. Conclusions
as B = 0.4, the temperature profile of the ring-coil model is closer to
that of the ‘‘solid’’ cylindrical model. This means that the ‘‘solid’’ Based on the ‘‘solid’’ cylindrical heat source model, an improved
cylindrical model can be accepted for the cases with dense ring analytical model of ring-coil heat source is established to better
coils. In addition, the temperature deviation between the two mod- illustrate the heat transfer process of the PGHE with spiral coils.
els shows a maximum value on the cylindrical surface of ring coils, In the ring-coil source model, the discontinuity of the heat source
where R = 1, and then rapidly attenuates at locations away from is considered, and the impact of the coil pitch is investigated by
the heat sources in the radial direction. This demonstrates that introducing more appropriate approximations of the real coil. In
the influence of pitch on the temperature response can be ne- this model, the buried coils are simplified as a number of separated
glected at the locations far away from the heat source in a short rings on the cylindrical surface with the pitch b. In order to inves-
term period. tigate the impacts of the ground surface and the limited length of
Fig. 8 compares the variations of the temperature rises on the the PGHEs, both the infinite and the finite-length heat source mod-
cylindrical surfaces of R = 1 for the two models with the time els are developed and their analytical solutions are derived accord-
elapses. Since the cylindrical heat source can be reckoned as a col- ing to the Green’s function theory and the image method. Both the
lection of numerous ring-coil heat sources with a same heating temperature distributions for the two models and the impact of the
intensity per length as the ring-coil source model, there is no tem- coil pitch on the thermal process are discussed respectively for
perature difference along the Z direction for this model. Conse- comparison based on the simulation results.
quently, Fig. 8 demonstrates that the temperature rise from the The simulation results give that the infinite ring-coil model may
ring-coil source is much larger in the vicinity of the ring sources be acceptable for engineering applications for short-time system
and slightly smaller near the middle points of the ring coils com- operation. However, for long-term duration, it is vital to consider
pared with those of the cylindrical model. However, in the design the heat conduction through the two ends of the pile, especially
of the PGHEs in GCHP systems, special attention should be paid for the pile with smaller depth-to-diameter ratio.
to the temperature rise on the pile wall, where is quite close to Comparisons between the cylindrical and ring-coil models indi-
the ring coils. As a result, a significant error may be caused by cate that the temperature rises of the ring-coil source model
the cylindrical model in calculating the temperature rise on the remarkably fluctuate around the temperature profiles of the cylin-
pile wall. Thus, it is essential to employ the ring-coil heat source drical model, with larger amplitude near the ring coil and smaller
model to design the PGHEs more accurately. one far away the ring coil. In conclusion, the ring-coil source model
of finite length can describe the heat conduction in the PGHEs with
spiral coils more accurately compared with both the infinite ring-
3.3. Comparisons of finite cylindrical and ring-coil source models coil and cylindrical source models.
It should be pointed out that the dimension of the coil pipe is
Fig. 9 compares the temperature distributions along the pile still neglected in the models and is assumed to be a ring-coil heat
depth obtained from the finite cylindrical and ring-coil source source. As a result, the temperature on the points, where the ring
models respectively. Basically, the temperature rises of the ring- sources locate, tends to infinity, which deviates from the realistic
coil source model fluctuate around the temperature profiles of conditions. Fortunately, it is still applicable in practice as the tem-
 P. Cui et al. / Applied Energy 88 (2011) 4113–4119 4119

perature rises on the pipe wall which is desirable for GHE design in [5] Pahud D, Fromentin A, Hubbuch M. Heat exchanger pile system for heating and
cooling at Zurich Airport. IEA Heat Pump Centre Newsletter. 1999;17(1):15–6.
the engineering applications is at certain distance away from the
[6] Laloui L, Nuth M, Vulliet L. Experimental and numerical investigations of the
ring source. Though the ring-coil source models can better describe behaviour of a heat exchanger pile. Int J Numer Anal Methods Geomech
the heat transfer process of the spiral coils and take the spiral bitch 2006;30(8):763–81.
into account, they still show a deviation from the practical heat [7] Yasuhiro Hamada, Hisashi Saitoh, Makoto Nakamura, Hideki Kubota, Kiyoshi
Ochifuji. Field performance of a PGHE system for space heating. Energy Build
transfer problem due to its assumption of spiral coil as ring-coil. 2007;39(5):517–24.
The application and experimental verification of the simulation [8] Sekine K, Ooka R, Yokoi M, Shiba Y, Hwang S. Development of a ground-source
model presented here will be addressed soon in a separate paper. heat pump system with ground heat exchanger utilizing the cast-in-place
concrete pile foundations of buildings. ASHRAE Trans 2007;113(PART
1):558–66.
Acknowledgements [9] Bose JE, Parker JD, McQuiston FC. Design/data manual for closed-loop ground-
coupled heat pump systems. Atlanta: American Society of Heating
Refrigerating and Air-Conditioning Engineers; 1985.
The research is supported jointly by a Grant from Shandong [10] Eskilson P. Thermal analysis of heat extraction boreholes. Doctoral Thesis,
Province Natural Science Foundation, China (Project No. Department of Mathematical Physics, University of Lund, Sweden; 1987.
ZR2010EQ006) and a Grant from Shandong Jianzhu University [11] Yavuzturk C, Spitler JD, Rees SJ. A transient two-dimensional finite volume
model for the simulation of vertical U-tube ground heat exchangers. ASHRAE
(010180902). Trans 1999;105(2):465–74.
[12] Zeng HY, Diao NR, Fang ZH. A finite line-source model for boreholes in
References geothermal heat exchangers. Heat Transfer Asian Res 2002;31(7):558–67.
[13] Kavanaugh SP. Simulation and experimental verification of vertical
groundcoupled heat pump systems. Doctoral Thesis Oklahoma State
[1] Spitler JD. Ground-source heat pump system research past, present and future.
University: Stillwater, Oklahoma; 1985.
HVAC and R Res 2005;11(2):165–7.
[14] Man Y, Yang HX, Diao NR, Liu JH, Fang ZH. A new model and analytical
[2] Yu MZ, Diao NR, Su DC, Fang ZH. A pilot project of the closed-loop ground-
solutions for borehole and pile ground heat exchangers. Int J Heat Mass Trans
source heat pump system in China. Proc IEA 7th Heat Pump Conf 2002:356–64.
2010;53:2593–601.
[3] Morino K, Oka T. Study on heat exchanged in soil by circulating water in a steel
[15] Jia L, Fang ZH. Advanced heat transfer. Beijing: China Higher Education Press;
pile. Energy Build 1994;21(1):65–78.
2008.
[4] Pahud D, Fromentin A, Hadorn JC. The duct ground heat storage model (DST)
[16] Beck J, Cole K, Haji-Sheikh A, Litkouhi B. Heat conduction using Green’s
for TRNSYS used for the simulation of heat exchanger piles. DGC-LASEN,
functions. Hemisphere 1992.
Lausanne; 1996.