Geothermics 81 (2019) 209–223

Contents lists available at ScienceDirect

Geothermics
journal homepage: www.elsevier.com/locate/geothermics

Pipe–pipe thermal interaction in a geothermal energy pile T

a,⁎ a b
Abubakar Kawuwa Sani , Rao Martand Singh , Cristina de Hollanda Cavalcanti Tsuha ,
Ignazio Cavarrettaa
a
Department of Civil and Environmental Engineering, Faculty of Engineering and Physical Sciences, University of Surrey, Guildford, GU2 7XH, UK
b
Department of Geotechnical Engineering, University of São Paulo at São Carlos, Av. Trabalhador Sãocarlense, 400, São Carlos, SP-13566-590, Brazil

A R T I C LE I N FO A B S T R A C T

Keywords: The use of energy loop(s), fitted into the structural foundation piles, also known as geothermal energy piles
Ground source heat pumps (GEPs) is on the rise. This dualizes the role of the piles in meeting the structural performance and the thermal
Thermal pile comfort demand of the overlying structure. Heat carrier fluid (HCF) is circulated through the loops, to extract or
Numerical modelling reject heat energy into the ground, during the space heating or cooling operation. However, this results in
Thermal interaction
thermal interaction between the inlet and outlet leg of the loop. This paper presents a numerical study to
investigate the pipe–pipe thermal interaction between the inlet and outlet loop–legs. It was found that factors
such as the number of loops, pipe location, soil and concrete thermal conductivity have a significant influence on
the magnitude of thermal interaction between inlet and outlet pipes. Similarly, it was found that the central steel
bar, used in contiguous flight auger (CFA) piles, contributes towards higher thermal interaction.

1. Introduction
The use of pile foundation elements as closed loop heat exchangers,
also known as geothermal energy piles (GEPs), thermal piles or pile
heat exchangers (PHEs), coupled to a heat pump has long been re-
cognized among the most energy-efficient means of achieving space
heating and cooling in residential and commercial buildings. The cou-
pling process connects the heat pump to the heat exchanger via the use
of high density poly-ethylene (HDPE) plastic pipes incorporated into
the GEP.
Within the GEP element, the HDPE pipes are often installed in a U-
shape manner and attached to the structural reinforcement cage prior
to lowering of the steel reinforcement cage into the bored hole and
concreting. Alternatively, in a contiguous flight auger (CFA) piles, the
HDPEs are bunched together around a central steel bar for rigidity and
plunged into the fresh concrete prior to concrete setting. Inside the
HDPEs, heat carrier fluid (HCF) which consists of water or water plus
antifreeze based solution is circulated using the heat pump to transfer
heat from the building to the ground via the GEPs during space cooling
or vice-versa in space heating operation. The combined system com-
prises the ground heat exchanger also known as the primary unit (high
density poly-ethylene (HDPE) loops and foundation element), heat
pump unit and the secondary unit (which transfer heat from the heat
pump to building) and are generally referred to as ground source heat
pump (GSHP) system, shown in Fig. 1.
In the GSHP system, the major part of the initial capital cost is at-
tributed to the GEP element (Lamarche et al., 2010). If the GEP is under
designed, the system will fail shortly in the early year(s) of operation.
On the other hand, if overdesigned, the system becomes unnecessarily
expensive and less energy efficient. Therefore, it is highly imperative to
ensure that the GEP is carefully designed, with the HDPE loops ap-
propriately situated, to ensure that a cost–effective and efficient unit,
which continuously exchange heat with the surrounding soil, is in-
stalled.
This is achieved by considering the heat transfer within the soil
domain as a transient phenomenon, due to its semi-infinite nature and
its thermal capacity. Whereas, within the GEP unit, a steady state ap-
proach is often assumed, with uniform radial and circumferential
temperature distribution, due to the finite dimension of the heat ex-
changer. However, numerous studies have shown that the distribution
of the radial and circumferential temperature in a GEP are not uniform
(Sani et al., 2018a; Loveridge and Powrie, 2014), as a result of factors
such as HDPE pipes configuration (shown in Fig. 2), number of loops
and their location (Park et al., 2013; Jalaluddin et al., 2011; Hamada
et al., 2007; Gao et al., 2008a, b; Mehrizi et al., 2016; Zarrella et al.,
2013a, b; Batini et al., 2015). Nonetheless, they are found to offer po-
sitive significance towards heat transfer if appropriate number are in-
stalled and suitably placed within the GEP.
In addition, numerous studies were carried out on improving the
heat transfer of the GEP by adding additives to the concrete mix to

⁎
Corresponding author.
E-mail address: a.sani@surrey.ac.uk (A.K. Sani).

https://doi.org/10.1016/j.geothermics.2019.05.004
Received 21 September 2018; Received in revised form 9 December 2018; Accepted 7 May 2019
0375-6505/ © 2019 Elsevier Ltd. All rights reserved.
A.K. Sani, et al. Geothermics 81 (2019) 209–223

Fig. 1. Ground source heat pump (GSHP) system.

Fig. 2. Different configurations of energy loops within a GEP.

increase its thermal conductivity (Remund, 1999; Lie and Kodur, 1996;
Xu and Chung, 2000a, b; Guo et al., 2010; Minto et al., 2016). Al-
though, this has its positive significance, however the structural in-
tegrity of the GEP should not be overlooked. In addition, a very high
thermal conductivity, would mean lower thermal resistance, and may
result in heat flow between the inlet and outlet leg(s) of the HDPE loop
(s) within the GEP also known as pipe-pipe thermal interaction. This
would affect the performance of the system, especially in a smaller
diameter GEP element with higher number of installed loops closely
situated to each other. This is especially common in CFA piles, where
the shank spacing between the inlet and outlet loops leg is negligible or
non-existent (Sani et al., 2018a, b). Hence, there is a need to investigate
the thermal interaction between HDPE pipes and how it varies with
higher number of loops.
This paper presents a numerical study carried out to investigate the
heat flow behaviour within a GEP using finite element method (FEM).
Firstly, it introduces the general background of the study, followed by a
review of the various available analytical approaches for determining
the resistance of a GEP. Secondly, it describes the numerical study in-
vestigating the thermal interaction between pipes and how it varies

210
A.K. Sani, et al. Geothermics 81 (2019) 209–223

considering factors such as number of pipes, pipe location, thermal be uniform at any point along the depth in the HDPE pipe. Thus, Rpconv
conductivity and specific heat capacity. Finally, the results are dis- is given by the following expression:
cussed, and the conclusions are outlined. 1
Rpconv =
2nπri hi (4)
2. Background
Where hi (W/m2 K) is the heat transfer coefficient determined using the
The heat flow (qp) within a GEP is governed by the temperature convection correlation and the Nusselt number (Nu) in Eq. (5). The
gradient and the thermal resistance property of the GEP material shown Nusselt number is expressed based on Dittus-Boelter equation
in Eq. (1). Nu λ fluid 0.023Re0.8 Pr0.35 λ fluid
hi = =
RGEP = ΔT/qp (1) 2ri 2ri (5)

where ΔT is the temperature difference between the HCF and the sur- where λfluid is the thermal conductivity of the HCF, Pr is the Prandtl
face of the GEP. RGEP is the thermal resistance. number (the ratio of the rate of viscous diffusion to thermal diffusion)
The thermal resistance of a GEP element (RGEP) is determined in a and Re is the Reynolds number (the ratio of the inertial forces to viscous
similar way to that of a borehole (Rb), however, the latter are non- forces for fluid flow).
mechanical load resisting elements and possesses greater length to The first term in Eq. (2) i.e. the concrete thermal resistance (Rc) is
diameter ratio compared to GEPs. difficult to evaluate and can be determined using various analytical
Owing to the smaller diameter in boreholes, heat transfer in the methods reported in the literature.
axial direction could potentially be significant and can be accounted for (a) Equivalent radius method
by using either effective mean fluid temperature or using effective One of the earliest methods of determining Rc assumes the U-tube to
borehole thermal resistance (Rb* ). Various analytical approaches to es- be a pipe of an equivalent radius (req ) having the same axis with the heat
timate Rb* , to account for the radial heat flow and the axial heat transfer exchanger. The concrete resistance is computed considering radial heat
that occurs between the inlet and outlet loops, have been proposed by conduction from the outer wall of the pipe to the borehole circumfer-
Lamarche et al. (2010), Zeng et al. (2002), Diao et al. (2004) and ential surface as given in Eq. (6).
Claesson and Hellström (2011). The complexity of these analytical 1
models varies as the number of incorporated loops increases. Thus,
Rc =
2πλ c ( )
ln rb req
(6)
making them time consuming and cumbersome to be accounted for
during the design process. Javed and Spitler (2016) showed that Rb* is Where λ c is the concrete thermal conductivity and rb is the radius of the
largely dependent on the borehole length, because it increases with the borehole. The req value can be computed using the expression proposed
increase in depth and decrease in the HCF flowrate. by Claesson and Dunand (1983) and used by Shonder and Beck (2000)
However, in a GEP element, the heat flow is assumed to occur shown in Eq. (7). However, the expression is quite simple and neglect
mainly in the radial direction. This assumption is reasonable and is the shank spacing (S) effect between the inlet and outlet U-tube pipes.
justified based on the findings of Singh et al. (2015a, b). They showed Other alternative solutions that account for the effect of shank spacing
that the heat transfer in a GEP element predominantly occurs in the were proposed by Bose et al. (1985) and Gu and O’Neal (1998) shown
radial plane. in Eqs. (8) and (9) respectively.
Similarly, Javed and Spitler (2016) studied the effect of varying the req = 2 ro (7)
borehole depth and the HCF flowrate on the effective borehole thermal
resistance. They showed that Rb and Rb* are equal for a 50 m deep req = ro S (8)
borehole. For a borehole with depth up to 100 m, Rb* was found to be
4% higher than Rb. Also, they showed that for a 200 m deep borehole req = 0.414 ro + 0.5 S (9)
with variable flowrates, Rb* was found to be 12% and 43% higher than
(b) Empirical method
Rb at 2 m3/hr and 1 m3/hr. Moreover, as the flowrates decreases from
Another method of calculating Rc of a GEP is through the use of
0.6 m3/hr to laminar flow i.e. flowrates lower than 0.2 m3/hr, the dis-
various empirical approaches available. It involves establishing an
crepancy between Rb* and Rb increases from 100% to over 500 times
empirical relationship derived from the use of curve fitting technique of
that of Rb values.
experimental and/or numerical data.
One of the popular approach of estimating Rc is that developed by
2.1. Review of available analytical approaches Paul (1996) and used by Remund (1999).
1
Different approaches for determining the thermal resistance of a Rc =
borehole element have been proposed. The same methods are used in Sb λ c (10)
determining GEP resistance (RGEP). Although the approaches vary, but The method is based on the shape factor (Sc) relationship de-
they all categorise the overall thermal resistance of a GEP (RGEP) into termined using empirical coefficients β0 and β1 which are back calcu-
three components namely: the convective resistance associated with the lated from experimental data.
HCF (Rpconv); the pipe conductive resistance (Rpcond); and the concrete
β1
conductive thermal resistance (Rc) given in Eq. (2). Sc = β0 rb ro
( ) (11)
RGEP = R c + Rpconv + Rpcond (2) The coefficients β0 and β1 for the three different configurations
The conductive thermal resistance through the HDPE pipe is com- (cases A–C), shown in Table 1, were investigated by Paul (1996). Case A
puted using the following equation: involves the inlet and outlet leg of the loop being in direct contact at the
centre of the borehole. Case B represents a case where the inlet and

Rpcond =
ln ( )ro
ri outlet legs are situated at equal distance from each other and the
2nπλp borehole wall. While case C considers a situation with maximum shank
(3)
spacing possible, where the U-tube legs are situated at opposite sides
Where n is the number of pipes, ri and ro are the pipe internal and outer touching the borehole edge.
radius, respectively. λp is the thermal conductivity of the pipe material. Similarly, Sharqawy et al. (2009) also developed an expression,
The radial temperature distribution in the HCF is usually assumed to presented in Eq. (12), to determine the borehole thermal resistance

211
A.K. Sani, et al. Geothermics 81 (2019) 209–223

Table 1 developed a complex algorithm which could be used in determining Rc

Borehole factor values for different U-tube configurations (after Remund Paul, for any number of pipes placed arbitrarily in the heat exchanger. The
1996). algorithm was developed in Fortran and Mathcad codes and was used in
Earth Energy Designer and GLHEPRO models (Javed and Spitler, 2016).
It is generally regarded as one of the most accurate ways of calculating
Rc (Lamarche et al., 2010; Loveridge and Powrie, 2014; Javed and
Spitler, 2016; Al-chalabi, 2013). The multi-pole solution for the case of
two symmetrical pipes is presented in Eq. (17).

⎡ ⎛ rb4 ⎞⎤
1 ⎢ r r
Rc = β + ln ⎛ b ⎞ + ln ⎛ b ⎞ + σln ⎜
⎜ ⎟
⎥
4 ⎟⎥
4 πλ c ⎢ ⎝ ro ⎠ ⎝S⎠
⎢
⎣
⎜ rb4 − S 2
⎝ ( ) ⎟
⎠⎥⎦

based on 2D finite element simulations. The authors claim that their

−
1 ⎢
⎡
⎢
ro2 2
S {1 − σ ( ( S4
4 )/r
4
b − ( ) )}
S
2
4 ⎤
⎥
⎥
expression is highly accurate within the range of dimensionless geo- 4 πλ c ⎢ 2
⎞⎫ ⎥
4
metrical parameters 2.5 ≤ db dp ≤ 7 and 0.2 ≤ S db ≤ 0.8.
⎣
(
⎢ 1+
β
1 −β + ) ( ) ⎧⎨⎩1 + σ ⎛⎝ (S r )/(r
ro2 2
S
4 4
b
4
b − ( ))
S
2
⎬⎥
⎠⎭ ⎦
1 ⎡ S d ⎤ (17)
Rc = −1.49 ⎛ ⎞ + 0.656 ln ⎜⎛ b ⎟⎞ + 0.436⎥
2 πλ c ⎢
⎜ ⎟

d
⎝ ⎠b d
⎝ p⎠ (12) Where β = 2πλ c Rp .
⎣ ⎦
Where db is the borehole diameter, S is the shank spacing and dp is the
2.2. Finite element modelling
pipe diameter.
Another best–fit expression of determining Rc has been proposed by
A pseudo 2D finite element modelling and analyses were carried out
Loveridge and Powrie (2014). The expression is given in Eq. (13).
using a numerical code known as COMPASS (Code for Modelling
1 Partially Saturated Soils). It was developed at Cardiff University, U.K.
Rc =
Sc λ c (13) and its theoretical formulations are based on mechanistic approach. It
has the capability of numerically solving a Thermo-Hydraulic-
The shape factor (Sc) expression shown in Eq. (14) was obtained
Mechanical and Chemical (THM-C) processes in a partially saturated
through sensitivity analyses of non–dimensional parameters rb ro and ( ) porous media. The extensive details of its theoretical formulations can
( ).
rb
cc be found in various publications (Thomas and He, 1997; Thomas and
A Sansom, 1995; Thomas and Li, 1996; Seetharam et al., 2007) and dis-
Sc = D E sertations (Sedighi, 2010; Vardon, 2009; Masum, 2012; Singh, 2007).
B ln ( ) + C ln ( ) + ( )
rb
ro
rb
cc
rb
ro + ( )
rb
cc +F (14) In this current study, only the thermal capability of the COMPASS
code was used. Thus, the governing equations for heat transfer ex-
Where cc is the concrete cover. The constants A, B, C, D, E and F are
pressed in terms of primary variables are reported here. The numerical
determined through curve fitting technique. Their values depend on
code computes the thermal problem by solving the heat transfer due to
number of loops and concrete–soil thermal conductivity ratio as pre-
liquid (i.e. pore water pressure (ul ) present in the soil pores) and that
sented in Table 2.
through solids (soil skeleton), due to temperature gradient (∇T ). The
(c) Multipole method
equation for heat transfer through soil media is computed numerically
Another expression for determining Rc is the line–source model
by the utilisation of the finite element approach to achieve spatial
presented by Hellström (1991), given in Eq. (15), and is used in the
discretisation, whereas the temporal discretisation of the solution is
Duct Storage model (DST) program (Lamarche et al., 2010) to simulate
calculated numerically via finite difference method (Vardon, 2009).
a vertical GEP and its surrounding soil.

⎡ 2.2.1. Heat transfer

1 ⎢ ⎛ rb ⎞ r ⎛ rp4 ⎞⎤
Rc = ln + ln ⎛ b ⎞ + σln ⎜ ⎥ With the effect of convection generally small, the most dominant
4 ⎟⎥
⎜ ⎟

4 πλ c ⎢ ⎝ ro ⎠ ⎝S⎠
⎢
⎣
⎜ rp4 − S 2
⎝ ( ) ⎟
⎠⎥⎦ (15)
heat transfer mechanism in soils is conduction, which depends on soil
porosity and water content. On the other hand, the effect of radiation is
Where rb ro and rb S are the most important parameters in Eq. (15), and σ usually negligible. However, in a soil of appreciable grain size, its effect
is determined using Eq. (16). becomes significant. Farouki (1981) reported that radiation effect could
result in an increase in heat transfer by up to 10% in gravels with a
λ c − λs
σ= particle size of 20 mm, compared to that in sand which is less than 1%.
λ c + λs (16)
The law of conservation of energy governing heat flow dictates that
In addition to the line–source equation, Bennet et al. (1987) the temporal derivative of the heat content (Hc ) is equal to the spatial
derivative of the heat flux (Q), mathematically expressed as:
Table 2
∂ ( Hc ∂V )
Curve fitting values for a single loop in a GEP (after Loveridge and Powrie = −∇Q (∂V )
∂t (18)
(2014)).
λc = λs λc = 2λs 2λc = λs Q = − λ s ∇T (19)

A 4.919 4.34 4.853 Where ∂t is change time, ∂V is the change in volume, λs is the soil
B 0.3549 0.317 0.345 thermal conductivity, ∇T is the temperature gradient.
C −0.07127 −0.001228 −0.1676
D −11.41 −10.18 −16.76
2.2.2. Model description
E −2.88 −2.953 −3.611
F 0.06819 −0.002101 0.1938 This section presents the description of the numerical models, the
initial and boundary conditions used for the different analyses.

212
A.K. Sani, et al.
Fig. 3. (a) Pile-only model geometry, and (b) pile-ground model geometry.
Fig. 4. Typical geometry of a rotary bored GEP with (a) 2–loops, (b) 3–loops and (c) 4–loops.
However, it should be noted that only heat transfer in the radial di-
rection was investigated and heat movement along the pile length was
not considered in this study.
(a) Pile-only model
A two–dimensional geometry of a single GEP shown in Fig. 3a was
set up in COMPASS. The model comprises a pile cross-section with a
dimeter of 600 mm and two pipes installed to represent the inlet and
outlet legs of a single U-loop. The modelled pipes have an outer and an
inner diameter of 32 and 28 mm, respectively. They are installed at the
edge of the GEP, with a concrete cover (cc) of 50 mm from the pile
outer surface. The inlet and outlet legs are separated by a shank spacing
(S) of 436 mm from the pipe-pipe surface.
The pile–only model is restricted only to the concrete GEP and the
HDPE pipes, i.e. the surrounding soil and the HCF were not modelled.
The analyses were carried out in order to compare the COMPASS results
with that of the available analytical approaches that do not take into
account the effect of soil thermal conductivity. Similarly, the convective
resistance of the HCF can easily be computed using Eq. (4), and their
values are significantly lower than Rc provided that a turbulent flow
regime is maintained. Thus, the effect of Rpconv was neglected in this
study.
The pile concrete surface was discretised using a uniform 4 noded
triangular mesh elements with an equal size of 10 mm. On the other
hand, the HDPE pipes were also discretised using 4 node triangular
elements however, with smaller elements size of 1 mm. At the regions
close to the pipes, 1 mm mesh sizes were applied at the pile-pipe in-
terface, and expand to 10 mm at other locations within the pile sur-
faces.
(b) Pile-ground model
In order to analyse the problem from a more realistic point of view,
the model in Fig. 3a was extended to include the soil surrounding the
213
Geothermics 81 (2019) 209–223
GEP cross–section. The GEP diameter of 600 mm was maintained as in
Fig. 3a, while the ground domain was extended outwards to a radial
distance of 5 m from the centre of the pile. The layout of the pile-ground
model domain is shown in Fig. 3b.
Furthermore, just as in the pile-only model case, the convective
resistance due to HCF was neglected in all the simulations carried out
using the pile-ground model.
The meshing described in the pile-only model, i.e. elements sizes of
10 mm and 1 mm on the pile and the pipe surfaces, were maintained
here. However, in the soil domain, the mesh size expands from the pile
surface towards the far field soil boundary. A 4 noded triangular mesh
of 20 mm size was applied at the pile surface and the elements size
increases to 500 mm at the far field boundary.
In addition to the two pipes accounted for in the pile-ground model
shown in Fig. 3b, other models were created to represent the cases
having higher number of loops incorporated in the pile cross–section.
Fig. 4a–c shows the pile geometry of a typical rotary bored pile with 4,
6 and 8 pipes installed respectively.
Furthermore, Fig. 5a and b illustrate respectively the GEP geometry
with 1–loop and 2–loops attached to a central steel bar installed at the
centre of the pile, which is a typical case in CFA piles. The central steel
diameter of 40 mm was chosen for the two models to realistically utilise
what is commonly used in the industry as reported in Loveridge and
Cecinato (2015).
Lastly, it is important to point out that the pile-only model and pile-
ground model were both validated by comparing the numerical results
with the available analytical approaches presented in this study.
Afterwards, the pile-ground model was used to investigate the pipe-pipe
thermal interaction that occurs in a typical GEP during heating opera-
tion.
A.K. Sani, et al.
Fig. 5. Geometry of a CFA pile, with (a) 1–loop and (b) 2–loops attached to a central steel.
2.2.3. Initial conditions
The initial conditions assumed in the two models, i.e. pile only
model and pile-ground model, corresponds to the typical values of a
GEP embedded in London clay. The initial temperature of the concrete
GEP, the HDPE pipes and the surrounding London clay were assumed to
be around 13.4 °C. This agrees with the all-year-round constant ground
temperature value measured at East London and reported by Loveridge
et al. (2013) during a thermal response test on a vertical borehole.
Furthermore, the initial degree of saturation of the London clay was
assumed to be 100% being a representative of the fact that the soil is
fully saturated (Loveridge et al., 2013). Additionally, the effect of the
interaction between the atmospheric air and the pile was neglected.
This is because piles and their surrounding soil are covered by buildings
thus are not exposed to the atmosphere.
2.2.4. Boundary conditions
The boundary conditions applied to the inlet and outlet pipes were
selected based on the data sheet of GSHP system specification obtained
from Kensa heat pumps, UK and Water Furnace international heat
pump manufacturers (waterfurnace, 2016; Heat, 2016).
A constant temperature of 308.15 K (35 °C) and 303.15 K (30 °C)
were applied to the internal surface of the inlet and outlet leg of the
HDPE pipes, respectively. This GSHP system mode of operation re-
presents the process of heat injection into the soil or summer mode to
provide space cooling. The same temperature conditions were applied,
to the inlet and outlet pipes, for both the pile–only model and the pile
and extended ground model, respectively.
In addition, a fixed temperature of 286.55 K (13.4 °C) was applied at
the outer surface of the concrete GEP in the pile–only model. This is to
allow all the temperature changes to be fully developed within the
concrete pile domain. However, in the pile-ground model, the fixed
temperature condition (i.e. 286.55 K) was applied at the outer surface
of the soil domain (i.e. at 5 m radial distance from the centre of GEP).
2.3. Material parameters
The materials used in this study are adopted from the values that
were measured and reported in literature. London clay was chosen as
the soil type surrounding the GEP. Table 3 presents the set of material
properties for the HDPE pipe, concrete GEP and the London clay.
3. Numerical analyses
This section presents the detailed description of the different si-
mulations conducted for both the pile–only model and the pile-ground
model. Firstly, the models were validated by comparing the numerical
results using analytical and empirical approaches. Afterwards, the pile-
214
Geothermics 81 (2019) 209–223
ground models were used to study the pipe-pipe thermal interaction
and how it varies with number of loops, thermal conductivity and pipe
location.
In addition, it should be noted that all the numerical simulations
were carried out for the duration of 6 months of continuous heat ap-
plication process. This is to represent a case where the GEP system was
installed in an equal space cooling (heat application to the ground) and
heating (ground heat extraction) dominated climate. However, this
study is limited to the ground heat application process, i.e. the ground
heat extraction was not investigated. This ensures the decoupling of the
thermal cyclic loading process, and allows for rigorous investigation of
the pipe-pipe thermal interaction characteristics in a GEP, with a focus
on factors that could significantly influence it.
Similarly, it should be understood that, unless otherwise stated, the
pile-ground model in Fig. 3b was used to carry out the numerical
analyses. This is to ensure that the effect of each of the factors (number
of loops, variation in thermal conductivity, loop location and specific
heat capacity) were individually studied to gain an appreciable un-
derstanding of their influence towards pipe-pipe thermal interaction.
3.1. Model validation
3.1.1. Pile–only model
The numerical simulation conducted using the pile–only model was
carried out by applying the described boundary conditions on to the
model. The analyses were carried out until a steady state condition was
reached within the concrete GEP cross–section. Additionally, a sensi-
tivity analysis was conducted by refining the size of the triangular
elements mesh to choose the best optimum mesh size. The final op-
timum mesh sizes adopted for the pipe and the GEP were described in
Section 2.2.2 of this work. The mesh was chosen when the results ob-
tained do not vary significantly after varying the elements size.
In order to validate the model, the results obtained using the op-
timum chosen mesh was utilised to compute the GEP thermal resistance
(RGEP). The result obtained from the numerical simulation was com-
pared with that obtained from the analytical approaches (described in
Section 2.1).
3.1.2. Pile-ground model
In the case of validating the pile-ground model, the geometry with
the two pipes (i.e. inlet leg and outlet leg) shown in Fig. 3b was used for
the analysis. A transient analysis was carried out for 6 months of con-
tinuous heating process. An initial time step of 100 s was found to yield
a converged result.
The mean temperature at the circumference of the concrete GEP
(TGEP) was used to calculate the shape factor (Sc) using Eq. (20) ex-
pressed as:
A.K. Sani, et al. Geothermics 81 (2019) 209–223

Table 3
Material parameters.
Parameter London Clay Concrete HDPE Pipe Steel Plastic bar

Hydraulic parameters
Saturated hydraulic conductivity, 5.8 × 10−11 (Hepburn, – – – –
ksat (m/s) 2013)
Degree of saturation, (%) 100 (Loveridge et al., 2013) – – – –
Thermal parameters
Thermal conductivity, λ (W/m K) 1.5 (Amis et al., 2008) 1.5 (Cecinato and 0.385 (Cecinato and 43 (Metals, 0.36 (DirectPlastics, 2008)
Loveridge, 2015) Loveridge, 2015) 2005)
Specific heat capacity, Cp (J/Kg K) 1820 (Cecinato and 1050 (Cecinato and 1465 (Gao et al., 2008a, b) 473 (Metals, 1500 (DirectPlastics, 2008)
Loveridge, 2015) Loveridge, 2015) 2005)
Other parameters
Density, ρd (Kg/m3) 1968 (Low et al., 2015) 2210 (Cecinato and 1100 (Gao et al., 2008a, b) 7801 (Metals, 1150 (DirectPlastics, 2008)
Loveridge, 2015) 2005)
Porosity, n 0.37 (Midttømme, 1998) 0.1 – –
Latent heat of vaporisation, L (J/Kg) 2,400,000
Henry's volumetric coefficient of 0.02
solubility, Hs
Specific gas constant 287.1

qp
Sc =
λ c (ΔT ) (20)

Tinlet − Toutlet
ΔT = TGPE − ⎛ ⎞
⎝ 2 ⎠ (21)

3.2. Investigating the effect of pipe–pipe heat flow

Heat flow characteristics that occur between the inlet and outlet leg
of a HDPE loop incorporated into a GEP were investigated. The tran-
sient thermal analysis was carried out for 6 months (180 days) of
continuous heat application process. In order to gain an appreciable
understanding into the pipe–pipe thermal interaction between the inlet
and outlet leg of the loop, the pile-ground model with a U–loop was
utilised for the analysis. The temperature distribution between the inlet
and outlet leg of the loop and how it varies with time were studied.
Furthermore, to investigate the effect of the difference in tempera-
ture between the inlet and outlet leg of the U–loop, on pipe–pipe heat
interaction, two sets of analyses were conducted. The first analysis in-
volved maintaining a temperature difference of 5 °C between the pipes
during the analysis i.e. an outlet and an inlet temperature of 30 and
35 °C, respectively. However, in the second analysis, the temperature at
the outlet leg of the loop was maintained at 30 °C while that at the inlet Fig. 6. GEP cross–section showing points A, B, C, D, E and F.
leg was increased to 45 °C. Therefore, in the second analysis a tem-
perature difference of 15 °C was imposed between the inlet and outlet temperature distribution at the pile surface and how it varies depending
leg of the loop. To study the results of the two distinct simulations, the on the number of loops are presented and discussed.
result of temperature evolution at some specific points (i.e. points A, B,
C, D, E and F) within and at the circumference of the GEP were obtained
and studied. The points are schematically shown in Fig. 6. 3.4. Effect of thermal conductivity variation

3.3. Effect number of installed loops
Different sets of numerical simulations were conducted to in-
vestigate the effect of number of installed loops on pipe–pipe thermal
interaction within a GEP. The main rational behind this investigation is
based on the conventional understanding that incorporating higher
number of loops into the GEP, allow greater amount of energy exchange
with the ground, owing to higher contact surface area of the pipes.
However, despite its appeal, incorporating additional number of
loops could potentially result in higher thermal interaction between the
different legs of the installed loops, especially when the loops are clo-
sely located to each other, as reported by Cecinato and Loveridge,
(2015). This effect was investigated by running a numerical simulation
for 6 months of continuous heating, using the different models shown in
Figs. 3b, & 4 a–c, respectively. The results were obtained in the form of
temperature evolution at points A–F shown in Fig. 6. Also, the
The variation in the thermal conductivity of the concrete GEP and
its surrounding soil could have a significant implication on the pipe–-
pipe thermal interaction that could occur in a conventional GEP. The
effect was studied by considering three different possible scenarios that
could occur with regards to a GEP. Firstly, a situation where the
thermal conductivity of the concrete GEP and the soil are equal.
Secondly, a case where the soil thermal conductivity is higher than that
of the concrete GEP, and thirdly a situation where the concrete con-
ductivity is higher than that of soil. The detailed ratio of the ground to
soil thermal conductivity used are shown in Table 4.
The numerical investigations were conducted for 6 months of con-
tinuous heating, using the model shown in Fig. 3b. The results were
obtained at points A, B and C.

215
A.K. Sani, et al.
Table 4
Possible cases of thermal conductivity.
Possible scenarios Thermal conductivity ratio (λc : λs)
λc = λs 1:1
λc = 0.5 λs 1:2
0.5 λc = λs 2:1
Table 5
Values of Specific heat capacities of soil and the concrete GEP (Cecinato and
Loveridge, 2015; Environment IESV, 2019).
Cases Cpc (J/kg K) Cps (J/kg K)
1 880 1820
2 880 1000
3 1000 1000
4 1000 1820
3.5. Effect of specific heat capacity (Cp)
This section describes the numerical investigations conducted to
study the effect of specific heat capacity of the GEP (Cpc) and sur-
rounding soil (Cps), on the pipe–pipe thermal interaction. Different
values of the Cpc and Cps were obtained from the literature and were
used in carrying out this investigation. In total, four different cases were
studied, case 1-4. The different cases with the respective values of the
Cp considered for the GEP and the soil are given in Table 5.
The numerical analyses were carried out using the model with a
single loop shown in Fig. 3b, by subjecting the GEP and its surrounding
soil to a continuous heating for 6 months. The results are reported and
discussed in Section 4.
3.6. Effect of pipe location
The location of where the loop(s) are installed within the GEP could
have a significant implication towards pipe–pipe thermal interaction. In
the case of rotary bored GEPs, i.e. where the steel reinforcement is
installed prior to the concreting process, the HDPE loop(s) could be
attached to the cage. Thus, allowing for greater shank spacing to be
achieved. In the case of CFA piles the loops are bunched together and
installed at the centre of the GEP after concreting but prior to the start
of the concrete setting and hardening process.
Although there is a higher likelihood of pipe–pipe thermal interac-
tion in the CFA GEPs, due to its smaller shank spacing. However, it is
also expected that rotary bored GEPs could potentially experience the
same effect, but at a lower magnitude than that of CFA, once the heat
injection/extraction process approaches a steady state.
To investigate this effect, in both rotary and CFA GEPs, the nu-
merical model shown in Figs. 3b and 5 a were used. The transient nu-
merical simulations were carried out for 6 months of continuous heat
injection process. Furthermore, in order to investigate the effect of in-
crease in number of installed loops on the pipe–pipe thermal interac-
tion, for the two different GEP types, more numerical simulations were
conducted. The models shown in Figs. 3b, 4 a, 5 a and b, representing
the case of single and double loops, were used for the different nu-
merical analyses.
The results were obtained in the form of the final temperature at-
tained after the 6 months of the transient heating analyses, taken at
some specific points (i.e. C, G H and I) within the GEP. These points are
schematically shown in Fig. 7. The subscripts “E and C” denotes the
location of the loop i.e. at the edge and the centre of the GEP.
3.7. Effect of central steel
In the CFA GEP construction technique, the loops are installed with
216
Geothermics 81 (2019) 209–223
the help of a central steel of adequate stiffness into the fresh concrete.
Hence, it is expected that the steel, because of its higher thermal
properties (outlined in Table 3), could greatly contribute to the pipe–-
pipe thermal interaction problem. In order to investigate that, two sets
of numerical analyses were conducted for the case where a single loop
and double loops are installed. In the first case, the properties of the
central steel bar are considered and represented in the model. While in
the second case, the central steel bar was replaced with a plastic bar of
adequate strength and poor thermal properties. The plastic bar was
made from Polyamide 66 which had an outer diameter of 40 mm
(DirectPlastics, 2008). Its thermal properties are given in Table 3.
The results were obtained at points C, G and H in the single loop
GEP model, whereas, in the double loop model, they were obtained at
points C, G and I, respectively as shown in Fig. 7.
3.8. Results & discussion
This section reports and discusses the results obtained from the
different numerical simulations and analytical approaches carried out.
Firstly, the result of the numerical modelling obtained using the pile
only model is compared with that of the different analytical solutions
reported. Afterwards, the numerical result obtained from the pile-
ground model were validated using the analytical approach reported by
(Loveridge and Powrie, 2014).
Furthermore, the results of the numerical modelling investigating
the influence of the different factors such as number and location of
pipes, thermal conductivity and specific heat capacity are reported and
discussed.
3.9. Model validation
3.9.1. Pile–only model
The results of the GEP thermal resistance (RGEP) for the different
analytical and empirical formulas outlined in Eqs. (6), (10), (12), (13),
(15) and (17) are reported and compared with the result obtained from
the COMPASS numerical simulation using the pile–only model as
shown in Fig. 8. The values of the RGEP obtained from the different
approaches vary in accuracy when compared to the COMPASS pi-
le–only model. All the methods provided RGEP, which are within the
range of 0–7% of the COMPASS model, except for the case proposed by
Paul (1996).
The Paul’s model gave a RGEP value that is 0.101 (m.K/W) higher
compared to the COMPASS model. This could be as a result of the
limitation attributed to the Paul’s approach i.e. it does not accurately
take into account the exact position of the installed loop. Additionally,
the large discrepancy could be due to the fact that case C shown in
Table 1, with the pipes touching the circumference of the GEP, was used
to compute the Paul’s equation. Meaning the concrete cover was not
accounted for, which is on the contrary compared to the other models.
Thus, making the Paul’s expression difficult to be employed for accu-
rately calculating RGEP.
On the other hand, the closest RGEP value was obtained using the
Bennet’s multi pole model, with a deviation of about –0.004 (m.K/W)
from the COMPASS model. This captures the powerful nature of the
multi pole expression in evaluating the GEP thermal resistance. The
next in agreement is the model proposed by Loveridge and Powrie, Eq.
(13), with a RGEP value that is 0.006 (m.K/W) higher than the COM-
PASS model. The models of Hellström and Shonder & Beck provided a
resistance that was 0.008 (m.K/W) higher, while Sharqawy’s expression
gave a RGEP value that was 0.11 (m.K/W) higher compared to that of
COMPASS.
3.9.2. Pile-ground model
The result of the shape factor (Sc) value determined, using Eq. (20),
was compared with the analytical equation proposed by Loveridge and
Powrie (2014) expressed in Eq. (14). The mesh was refined until the
A.K. Sani, et al. Geothermics 81 (2019) 209–223

Fig. 7. GEP cross–section showing points C, G, H and I.

respectively.
The results show the capability of the Loveridge and Powrie’s
(Loveridge and Powrie, 2014) expression, in accurately determining
RGEP, and also in taking into account the variable circumferential
temperature distribution around the GEP perimeter.

(ScCOMPASS − Scanalytical )2
RMSE =
ScCOMPASS 2 (22)

|(ScCOMPASS − Scanalytical )|
Errmax =
ScCOMPASS (23)

3.10. Pipe–pipe heat flow

Fig. 8. GEP resistance computed for COMPASS and the different analytical
approaches.
The results of the radial heat distribution between the inlet and
outlet leg of the loop at different simulation time, i.e. 100 s (initial time-
Table 6 step), 4 h, 1, 5 30 and 180 days, respectively, are shown in Fig. 9.
Sc, RMSE and Errmax Results. Regions closer to the pipes witnessed greater temperature change over
Models Sc RMSE Errmax time and the effect diminishes towards the GEP centre where the lowest
effect was observed.
Loveridge and Powrie (2014) 4.986253942 0.0000093120702 −0.00305
At the start of the test (first 4 h), no significant thermal exchange
COMPASS 4.971084337
was observed, at the centre of the heat flow path (point C), between the
two pipes. However, at a distance of 100 mm, from the faces of inlet and
results of the Sc computed using Eqs. (20) and (14) do not vary by more outlet pipes, about a 4 and 3 °C increase in temperature above the initial
than 0.3%, as shown in Table 6. To further evaluate the accuracy of the state were observed.
numerical model result, the root mean square error (RMSE) and the As the test continues, a significant increase in temperature of about
maximum error (Errmax) were determined using Eqs. (22) and (23), 7 and 10 °C were observed, after 1 and 5 days of operation, respectively,
at 100 mm from the surfaces of the inlet and outlet loop legs. Whereas,

217
A.K. Sani, et al.
Fig. 9. Pipe-pipe heat flow at different simulation time.
at point C, a lower temperature of about 6 and 9 °C were observed, after
the 1 and 5 days operation, respectively.
Furthermore, the rate of temperature increase reduces after 5 days
of simulation as temperature at all points approach steady state
(Fig. 10). Steady state is defined here as when the changes in tem-
perature at all points have reached about 98%. After 30 and 180 days of
operation, a slight increase in temperature of about 2 and 3.5 °C (above
the values monitored after 5 days) were observed, at all the points,
along the heat flow path, respectively.
In addition, the contribution of the inlet and outlet loop–leg towards
thermal interaction were calculated by computing the thermal re-
sistance between the two legs, using the three-resistance model
(Lamarche et al., 2010). The thermal flux contribution from each leg
was obtained as the thermal potential divided by the computed re-
sistance (Myers, 1998). It was found that the heat flowing from inlet to
Fig. 10. Temperature evolution (a) at points A, B, and C, (b) at points D, E and F
with 35 and 45 °C inlet temperature.
218
Geothermics 81 (2019) 209–223
outlet leg was 4.83 (W/m). Whereas, the outlet loop–leg was calculated
to be contributing a heat flux of 1.4 (W/m) towards the inlet leg.
The temperature difference between the inlet and outlet loop–leg
have significant impact on the thermal interaction characteristics and
the time to achieve steady state (Fig. 10a & b). Increasing the tem-
perature difference from 5 to 15 °C, increased the heat flux flow by 0.58
(W/m). Thus, contributing to a higher heat flow towards the outlet
pipe, but at a far lower rate considering the large temperature gradient.
However, during the initial stages of the simulations, it was ob-
served that heat was flowing from the outlet leg towards the inlet
loop–leg. But after steady state has been reached, the temperature at
point C was found to be higher than 30 °C (outlet pipe temperature).
Hence, no heat was flowing from the outlet leg towards the inlet leg
after 180 days of operation.
Additionally, the time to achieve steady state at the points within
the GEP increases as the inlet and outlet temperature difference in-
creases. It was observed that the time increases from about 8–15 h at
points A & B, while at point C, it increases from 2 to 3.5 days, respec-
tively. However, at the GEP circumference, it increases from about 15 h
to 2 days (at points E & F) and from about 4–8 days at point D, as the
difference in temperature increases, respectively.
3.11. Influence of number of installed loops
The results of the effect of incorporating higher number of loops (1,
2, 3 and 4 loops), on thermal interaction and time to achieve steady
state, within the GEP are shown in Fig. 11a–f. The thermal interaction
between the inlet and outlet pipes could be related to the temperature
evolution at some points within and at the GEP circumference.
Inside the GEP (points A, B and C), increasing the number of in-
stalled loops from 1–loop to 2–loops, results in significant increase in
temperature by up to 4.5 °C after 180 days simulation. However, the
rate of temperature increase reduces, to 1.2 °C and 0.5 °C, as the number
of installed loops increases from 2–loops to 3–loops and from 3–loops to
4–loops, respectively (Fig. 11a–c). Whereas, at the GEP perimeter
(points D, E and F), a temperature increase of 3.9, 1.3 and 2 °C were
observed as the number of loops increases from 1–loop to 2–loops,
2–loops to 3–loops, and 3–loops to 4–loops (Fig. 11d–f).
Furthermore, the time to achieve steady state condition at all the
points considered within and at the perimeter of the GEP varies as the
number of installed loops increases (Fig. 11a–f). The duration decreases
from about 5 days to around 3 days, as the number of installed pipes
increases from 1–loop to 4–loops, for the 600 mm GEP diameter con-
sidered. However, the duration could differ as the diameter of the GEP
increases or decreases (Loveridge and Powrie, 2014).
Contrary to the assumptions made in simplistic design of GEPs, and
in some approaches of computing the GEP thermal resistance, the cir-
cumferential temperature is not constant. The distribution of tem-
perature along the perimeter of the GEP depends on the location of the
pipe (either at centre or edge), its function i.e. either inlet or outlet leg,
and on the number of installed loops. The distribution of temperature
around the GEP circumference, for 1, 2, 3 and 4 number of installed
loops, is shown in Fig. 12. Equal values of λc and λs were used for the
case considered. However, the variation in soil and concrete thermal
conductivity could have an impact on the magnitude of temperature
distribution around the GEP circumference.
3.12. Effect of thermal conductivity
The results of the numerical simulations carried out to investigate
the influence of λc and λs variation on pipe–pipe thermal interaction are
shown in Fig. 13. This figure shows the temperature evolution at points
A, B and C within the GEP for different cases of λc and λs. The tem-
perature at all the points within the GEP varies as the values of λc and λs
differ.
For the scenario where the soil thermal conductivity is higher than
A.K. Sani, et al.
Fig. 11. Temperature evolution for 1 loop, 2, 3 and 4 loops at points A, B, C, D, E and F.
Fig. 12. Temperature distribution around the GEP circumference.
that of concrete (λc = 0.5λs), the temperature increase at all locations
within and around the GEP circumference is minimal, and the heat
diffuses easily into the surrounding soil. Consequently, resulting in
lesser thermal interaction between the installed loop(s), and lower
magnitude of circumferential temperature distribution around the GEP,
compared to either of the cases when λc = λs (shown in Fig. 12) or
0.5λc = λs, respectively.
However, in a case where the concrete conductivity is higher (0.5λc
= λs), the surrounding soil creates a form of thermal insulation barrier
around the GEP. This results in higher temperature increase at all
points, and subsequently leading to greater thermal interaction be-
tween the inlet and outlet legs of the loop(s) than in the other two
scenarios. Similarly, it results in higher magnitude of circumferential
219
Geothermics 81 (2019) 209–223
Fig. 13. Temperature evolution at points A, B and C for different thermal
conductivity cases.
temperature distribution around the GEP in comparison to the case
shown in Fig. 12.
Furthermore, the time to achieve steady state condition depends on
both the thermal conductivity of the soil and the concrete GEP. The
time increases from just under about 3 days when λc = 0.5λs to around
4 and 5 days under the scenarios where λc = λs and 0.5λc = λs, re-
spectively.
3.13. Effect of specific heat capacity (Cp)
The results of the analyses conducted investigating the influence of
Cp on heat built-up within a GEP, are reported in Fig. 14, in terms of
temperature evolution at points A, B and C for the four cases
A.K. Sani, et al.
Fig. 14. Temperature evolution at points A, B and C for different cases of
specific heat capacity.
considered.
It was observed that Case–1 and Case–4, with Cps value of 1820 (J/
kg K), produced approximately the same results. Similarly, the results of
Case–2 and Case–3, having a Cps value of 1000 (J/kg K), were found to
be equal. The cases with the higher value of Cps = 1820 (J/kg K) re-
sulted in temperature evolution values that were lower by about 0.7 °C,
after the first day of operation, in comparison to the cases with the
lower value of Cps = 1000 (J/kg K).
In addition, it can be said that varying the Cpc from 880 to 1000 (J/
kg K) does not have any significant influence on the magnitude of
temperature evolution at all the points considered within the GEP.
Perhaps this is because of the fact that the range of the Cpc values
considered do not largely differ. Nonetheless, it can be deduced that
small variation in Cpc will not have any noticeable effect on the pi-
pe–pipe thermal interaction.
3.14. Effect of pipe location
The result of the analyses investigating the influence of pipe loca-
tion on pipe–pipe thermal interaction is shown in Fig. 15. The figure
shows a temperature distribution within the GEP, after 180 days of
continues heating operation, for a 1–loop and 2–loops arrangements
installed at the centre and edge of the pile. For all the simulations, a
constant temperature of 35 and 30 °C were imposed at the inlet and
outlet leg throughout the simulation time.
For the case with 1–loop installed, the temperature at the GEP
centre was found to be 31.5 °C when the pipes are at the centre, com-
pared to 26 °C when installed at the pile edge (Fig. 15a & b). Similarly,
for the case with 2–loops, the temperature at the centre of the pile were
found to be 32.8 and 30 °C, when the pipes were installed at the centre
and at the edge (Fig. 15c & d), respectively.
Thus, it can be deduced that having the loop(s) installed at the
centre, i.e. with smaller shank spacing, could eventually result in higher
thermal interaction and potentially leading to thermal short circuiting
of the system (Fig. 15a & c). This should be avoided by ensuring that
pipes are not bunched around a central steel, but rather are provided
with spacers to allow some space between the pipes in the case of CFA
piles. This will improve the overall system performance.
In addition, it was observed that the circumferential temperature
distribution around the GEP is non–uniform and the magnitude is
greater when the loop(s) are installed near the GEP edge. However, the
magnitude of the temperature developed around the GEP decreases and
its distribution becomes uniform when the pipes are at the centre. Thus,
resulting in uniform temperature changes in the soil surrounding the
pile.
3.15. Effect of central steel
The steel installed at the centre of a CFA pile for rigidity, shown in
220
Geothermics 81 (2019) 209–223
Fig. 15, has an influence on the heat transfer between the inlet and
outlet pipes. The results of this effect are shown in Fig. 16, for the cases
with 1–loop and 2–loops, installed with a steel and plastic bar at the
centre of the GEP. It was observed that the presence of the central steel
decreases the temperature difference between the pipes surfaces and
the centre of the GEP.
For the case with 1–loop, the temperature at the outlet leg (point G),
GEP centre (point C) and inlet leg (point H) were found to be 31.05,
31.26 and 31.85 °C, respectively when the central steel was installed in
the GEP. However, when the steel was replaced with a plastic bar, the
temperature at all the points decreases to 29.95, 31.00 and 34.15 °C,
correspondingly (Fig. 16a).
Similarly, in the situation with 2–loops, the temperature at the
outlet (point I), centre (point C) and inlet (point H) were found to be
31.95, 32.35 and 32.65 °C, when the central steel effect was accounted
for, in comparison to the temperature of 30.55, 32.35 and 34.25 °C with
the plastic bar, respectively (Fig. 16b).
Thus, it can be deduced that the difference in temperature between
the inlet and outlet pipes decreased from 4.2 °C (with the central plastic
bar) to 0.8 °C when the steel bar effect was considered for a 1–loop case.
Similarly, for 2–loops, the difference decreases from 3.7 °C to 0.7 °C for
the cases with the plastic bar and steel bar effect, respectively.
Therefore, it is recommended that hard plastics of adequate strength
should be used rather than the steel bar to improve the CFA GEPs
performance.
3.16. Influence of the various factors on the surrounding soil
The results of the effect of the different factors studied on the
temperature distribution in the soil surrounding the GEP are reported in
Fig. 17.
It was observed that placing the loops at the GEP centre results in
lower magnitude of temperature that develops in the soil surrounding
the pile. Similarly, the temperature magnitude increases with the
number of installed loops and their close proximity to the GEP edge.
Additionally, the region that is affected by the radial temperature dis-
tribution within the soil spans out to about 4.1 m away from the GEP
surface, in all the cases considered (Fig. 17a).
Furthermore, the value of λc and λs have significant influence on the
magnitude of temperature developed at the GEP soil interface. The
magnitude increases from 25 °C when 0.5λs = λc to about 30 °C when λs
= 0.5λc, respectively. Similarly, the thermally active region in all the
cases is within a radial distance of about 4.1 m from the pile (Fig. 17b).
4. Conclusion
This study investigates the 2D radial pipe–pipe thermal interaction
between the installed loops and how it varies based on some factors
including number of loops, variation in thermal conductivity and loops
location. It was found that:
• The analytical solutions proposed by Bennet et al. (1987) and
Loveridge and Powrie (2014) were found to produce the most ac-
curate results for the GEP thermal resistance with an error of about
+0.005 (m.K/W) in comparison to the COMPASS model. Whereas,
the Paul (1996) model gave the most divergent result, with a RGEP
value of +0.101 (m.K/W) greater than the COMPASS results, for the
pile only model.
• The results of the shape factor calculated using the equation pro-
posed by Loveridge and Powrie (2014) do not vary by more than
0.3% in comparison to the COMPASS result, for the case of pile-
ground model.
• Heat exchange between the inlet and outlet loop leg becomes sig-
nificant, after a steady state condition is achieved at all points, in the
GEP in a duration of 3–5 days of continues operation. This resulted
in pipe-pipe thermal interaction with 5.41 (W/m) flowing from the
A.K. Sani, et al.
Fig. 15. Temperature distribution (in °C), for 1–loop and 2–loops, at the centre and edge within the GEP.
Fig. 16. Effect of Central steel in CFA piles for (a) 1–loop and (b) 2–loops.
inlet pipe towards the outlet pipe, and 1.37 (W/m) of heat flowing
towards the inlet pipe.
• The number of installed loops results in higher magnitude of tem-
perature development in the GEP, and consequently leading to
221
Geothermics 81 (2019) 209–223
Fig. 17. Temperature distribution in soil for (a) different number of loops and
loop-location, and (b) different thermal conductivity values.
higher thermal interaction between the loops. Also, the number of
installed loops does have a significant influence on the temperature
distribution around the circumference of the pile. Moreover, the
time to achieve steady state decreases from about 5 days to around 3
A.K. Sani, et al.
days as the number of loops increases from 1 to 4.
• Lower magnitudes of temperature values were observed at all points
in the pile when λc = 0.5λs, in comparison to the case where 0.5λc
= λs. Thus, it can be said that higher thermal interaction and larger
magnitude of circumferential temperature distribution would occur
in the latter case.
•
The range of Cp values considered, for the concrete GEP and the soil,
do not have any major implication on the temperature development
and thermal interaction within and around the GEP.
•
The pipes installed at the GEP centre results in higher thermal in-
teraction between the inlet and outlet loop legs, and the pipe–pipe
thermal interaction decreases as the pipes moves closer to the GEP
edge. Similarly, the magnitude of circumferential temperature dis-
tribution around the GEP is greater and non-uniform when the pipes
are located closer to the edge, and the effect diminishes as the pipes
moves towards the GEP centre.
• The presence of central steel bar in CFA piles lead to higher thermal
interaction between the installed loops because of the high λ value
of the steel. Thus, it is recommended that a different material (say
plastic bar) of adequate strength with lower thermal conductivity
should be used rather than steel. Similarly, the use of spacers should
be adopted to reduce the radial thermal interaction.
• Higher magnitude of temperature develops in the soil when the
location of the installed loop is closer to the GEP edge, and its effect
increases with the increase in number of loops. Equally, greater
values of λc results in higher temperature development at the
soil–GEP interface compared to having a higher λs. Finally, the
thermally active region for a 600-mm diameter GEP, with 1–4 loops
installed at the edge and centre, was found to be within 4.1 m from
the GEP surface. However, this would differ in situations with
groundwater flow, unsaturated soils and where the soil conductivity
is higher.
References
Al-chalabi, R., 2013. Thermal Resistance of U-tube Borehole Heat Exchanger System:
Numerical Study. The University of Manchester.
Amis, T., Bourne-Webb, P.J., Davidson, C., Amatya, B., Soga, K., 2008. The effects of
heating and cooling energy piles under working load at Lambeth College, UK.
Proceedings of the 33rd Annual and 11th International Conference on Deep
Foundations. pp. 10.
Batini, N., Rotta Loria, A.F., Conti, P., Testi, D., Grassi, W., Laloui, L., 2015. Energy and
geotechnical behaviour of energy piles for different design solutions. Appl. Therm.
Eng. 86, 199–213. https://doi.org/10.1016/j.applthermaleng.2015.04.050.
Bennet, J., Claesson, J., Hellström, G., 1987. Multipole method to compute the conductive
heat flows to and between pipes in a composite cylinder. Notes Heat Transf.
Bose, J.E., Parker, J.D., McQuiston, F.C., 1985. American Society of Heating R and A-CE.
Design/data Manual for Closed-loop Ground-coupled Heat Pump Systems. American
Society of Heating, Refrigerating, and Air-Conditioning Engineers, Atlanta, GA.
Cecinato, F., Loveridge, F., 2015. Influences on the thermal efficiency of energy piles.
Energy 82, 1021–1033. https://doi.org/10.1016/j.energy.2015.02.001.
Claesson, J., Dunand, A., 1983. Heat Extraction From the Ground by Horizontal Pipes: A
Mathematical Analysis.
Claesson, J., Hellström, G., 2011. Multipole method to calculate borehole thermal re-
sistances in a borehole heat exchanger. HVAC R Res. 17, 895–911. https://doi.org/
10.1080/10789669.2011.609927.
Diao, N.R.R., Zeng, H.Y.Y., Fang, Z.H.H., 2004. Improvement in modeling of heat transfer
in vertical ground heat exchangers. HVACR Res. 10, 459–470. https://doi.org/10.
1080/10789669.2004.10391114.
DirectPlastics, 2008. Polyamide Material Datasheet. 2560889. https://www.dir-
ectplastics.co.uk/pdf/datasheets/Nylon 66 Black Data Sheet.pdf (accessed December
1, 2017).
Environment IESV. Apache-Tables User Guide n.d.
Farouki, O.T., 1981. Therm. Prop. Soils 136.
Gao, J., Zhang, X., Liu, J., Li, K., Yang, J., 2008a. Numerical and experimental assessment
of thermal performance of vertical energy piles: an application. Appl. Energy 85,
901–910. https://doi.org/10.1016/j.apenergy.2008.02.010.
Gao, J., Zhang, X., Liu, J., Li, K.S., Yang, J., 2008b. Thermal performance and ground
temperature of vertical pile-foundation heat exchangers: a case study. Appl. Therm.
Eng. 28, 2295–2304. https://doi.org/10.1016/j.applthermaleng.2008.01.013.
Gu, Y., O’Neal, D.L., 1998. Development of an equivalent diameter expression for vertical
U-tubes used in ground-coupled heat pumps. ASHRAE Trans. 104, 347–355.
Guo, C., Zhu, J., Zhou, W., Chen, W., 2010. Fabrication and thermal properties of a new
heat storage concrete material. J. Wuhan Univ. Technol.—Mater. Sci. Ed. 25,
628–630. https://doi.org/10.1007/s11595-010-0058-3.
222
Geothermics 81 (2019) 209–223
Hamada, Y., Saitoh, H., Nakamura, M., Kubota, H., Ochifuji, K., 2007. Field performance
of an energy pile system for space heating. Energy Build. 39, 517–524. https://doi.
org/10.1016/j.enbuild.2006.09.006.
Heat, K., 2016. Single Compressor Compact Heat Pump Installation and Commissioning
Manual.
Hellström, G., 1991. Ground Heat Storage: Thermal Analyses of Duct Storage Systems.
Lund University, pp. 310.
Hepburn, B.D.P., 2013. An Investigation of the Behaviour of the Ground in Response to
Energy Extraction.
Jalaluddin, Miyara, A., Tsubaki, K., Inoue, S., Yoshida, K., 2011. Experimental study of
several types of ground heat exchanger using a steel pile foundation. Renew. Energy
36, 764–771. https://doi.org/10.1016/j.renene.2010.08.011.
Javed, S., Spitler, J.D., 2016. Calculation of Borehole Thermal Resistance, vol. 1https://
doi.org/10.1016/B978-0-08-100311-4.00003-0.
Lamarche, L., Kajl, S., Beauchamp, B., 2010. Geothermics a review of methods to evaluate
borehole thermal resistances in geothermal heat-pump systems. Geothermics 39,
187–200. https://doi.org/10.1016/j.geothermics.2010.03.003.
Lie, T.T., Kodur, V.K.R., 1996. Thermal and mechanical properties of steel-fibre-re-
inforced concrete at elevated temperatures. Can. J. Civ. Eng. 23, 511–517. https://
doi.org/10.1139/l96-055.
Loveridge, F., Cecinato, F., 2015. Thermal performance of thermo-active CFA piles.
Environ. Geotechnol. 1–27.
Loveridge, F., Powrie, W., 2014. 2D thermal resistance of pile heat exchangers.
Geothermics 50, 122–135. https://doi.org/10.1016/j.geothermics.2013.09.015.
Loveridge, F., Holmes, G., Roberts, T., Powrie, W., 2013. Thermal response testing
through the Chalk aquifer in London, UK. Proceedings of the ICE - Geotechnical
Engineering 166, 197–210. https://doi.org/10.1680/geng.12.00037.
Low, J.E., Loveridge, F., Powrie, W., Nicholson, D., 2015. A comparison of laboratory and
in situ methods to determine soil thermal conductivity for energy foundations and
other ground heat exchanger applications. Acta Geotechnol. 10, 209–218. https://
doi.org/10.1007/s11440-014-0333-0.
Masum, S.Al., 2012. Modelling of Reactive Gas Transport in Unsaturated Soil. A Coupled
Thermo-hydro-chemical-mechanical Approach.
Mehrizi, A.A., Porkhial, S., Bezyan, B., Lotfizadeh, H., 2016. Energy pile foundation si-
mulation for different configurations of ground source heat exchanger. Int. Commun.
Heat Mass Transf. 70, 105–114. https://doi.org/10.1016/j.icheatmasstransfer.2015.
12.001.
Metals, A.S., 2005. Atlas Specialty Metals The Atlas Specialty Metals Technical Handbook
of Bar Products.
Midttømme, K., 1998. Thermal Conductivity of Selected Claystones and Mudstones from
England. Clay Miner. 33, 131–145. https://doi.org/10.1180/claymin.1998.033.1.12.
Minto, A., Leung, A.K., Vitali, D., Knappett, J.A., 2016. Thermomechanical Properties of a
New Small-scale Reinforced Concrete Thermo-active Pile for Centrifuge Testing. pp.
37–44.
Myers, G.E., 1998. Analytical Methods in Conduction Heat Transfer. AMCHT
Publications.
Park, H., Lee, S.R., Yoon, S., Choi, J.C., 2013. Evaluation of thermal response and per-
formance of PHC energy pile: field experiments and numerical simulation. Appl.
Energy 103, 12–24. https://doi.org/10.1016/j.apenergy.2012.10.012.
Paul, N.D., 1996. The Effect of Grout Thermal Conductivity on Vertical Geothermal Heat
Exchanger Design and Performance. Mechanical Engineering Department, South
Dakota State University.
Remund, C., 1999. Borehole Thermal Resistance: Laboratory and Field Studies.
Sani, A.K., Singh, R.M., Cavarretta, I., Tsuha C de, H.C., Bhattacharya, S., 2018a. Inlet &
outlet pipe heat interaction in a contiguous flight Auger (CFA) pile. In: Ferrari, A.,
Laloui, L. (Eds.), Energy Geotechnics. Springer, pp. 111–122.
Sani, A.K., Martand, R., Ignazio, S., 2018b. Numerical investigation on the performance of
geothermal energy contiguous flight auger (CFA) pile. Cui, L., Bhattacharya, S.,
Nikitas, G. (Eds.), 15th BGA Young Geotechnical Engineers’ Symposium 69–70.
Sedighi, M., 2010. An Investigation of Hydro-geochemical Processes in Coupled Thermal,
Hydraulic, Chemical and Mechanical Behaviours of Unsaturated Soils.
Seetharam, S.C., Thomas, H.R., Cleall, P.J., 2007. Coupled thermo/hydro/chemical/me-
chanical model for unsaturated soils—numerical algorithm. Int. J. Numer. Methods
Eng. 70, 1480–1511. https://doi.org/10.1002/nme.
Sharqawy, M.H., Mokheimer, E.M., Badr, H.M., 2009. Effective pipe-to-borehole thermal
resistance for vertical ground heat exchangers. Geothermics 38, 271–277. https://
doi.org/10.1016/j.geothermics.2009.02.001.
Shonder, J.A., Beck, J.V., 2000. Field Test of a New Method for Determining Soil
Formation Thermal Conductivity and Borehole Resistance, United States. American
Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc., Atlanta,
GA (US).
Singh, R.M., 2007. An Experimental and Numerical Investigation of Heat and Mass
Movement in Unsaturated Clays. Cardiff University.
Singh, R.M., Bouazza, A., Wang, B., 2015a. Near-field ground thermal response to heating
of a geothermal energy pile: observations from a field test. Soils Found. 55,
1412–1426.
Singh, R.M., Bouazza, A., Wang, B., Haberfield, C.H., Baycan, S., Carden, Y., 2015 2015b.
Thermal and thermo-mechanical response of a geothermal energy pile. World
Geotherm. Congress 7.
Thomas, H.R., He, Y., 1997. A coupled heat–moisture transfer theory for deformable
unsaturated soil and its algorithmic implementation. Int. J. Numer. Methods Eng. 40,
3421–3441. https://doi.org/10.1002/(sici)1097-0207(19970930)40:18<3421::aid-
nme220>3.0.co;2-c.
Thomas, H.R., Li, C.L.W., 1996. A parallel computing solution of heat and moisture
transfer in unsaturated soil. Int. J. Numer. Methods Eng. 39, 3793–3808.
Thomas, H.R., Sansom, M.R., 1995. Fully coupled analysis of heat moisture and air
A.K. Sani, et al.
transfer in unsaturated soil. J. Eng. Mech. 121, 392–405. https://doi.org/10.1680/
geot.1995.45.4.677.
Vardon, P.J., 2009. Three-dimensional Numerical Investigation of the Thermo-hydro-
mechanical Behaviour of a Large-scale Prototype Repository.
waterfurnace, 2016. ENVISION2 NKW 50Hz Reversible CHILLER Specification Catalog.
Xu, Y., Chung, D.D.L., 2000a. Effect of sand addition on the specific heat and thermal
conductivity of cement. Cem. Concr. Res. 30, 59–61. https://doi.org/10.1016/S0008-
8846(99)00206-9.
Xu, Y., Chung, D.D.L., 2000b. Cement of high specific heat and high thermal conductivity,
obtained by using silane and silica fume as admixtures. Cem. Concr. Res. 30,
1175–1178. https://doi.org/10.1016/S0008-8846(00)00296-9.
223
Geothermics 81 (2019) 209–223
Zarrella, A., Capozza, A., De Carli, M., 2013a. Performance analysis of short helical
borehole heat exchangers via integrated modelling of a borefield and a heat pump: a
case study. Appl. Therm. Eng. 61, 36–47. https://doi.org/10.1016/j.applthermaleng.
2013.07.021.
Zarrella, A., De Carli, M., Galgaro, A., 2013b. Thermal performance of two types of energy
foundation pile: helical pipe and triple U-tube. Appl. Therm. Eng. 61, 301–310.
https://doi.org/10.1016/j.applthermaleng.2013.08.011.
Zeng, H.Y.Y., Diao, N.R.R., Fang, Z.H.H., 2002. A finite line-source model for boreholes in
geothermal heat exchangers. Heat Transf.—Asian Res. 31, 558–567. https://doi.org/
10.1002/htj.10057.