Heat

TIG 6.
2: Final Report
Feasibility of Underground Cooling for

Geothermal Power Cycles
A/Prof. Bassam Dally, Prof. Gus Nathan

Mr Sean Watson, Mr Andrew Heath, Dr Carl Howard,
and Dr Peter Ashman
May 2009
This project is supported by the Primary Industry and Resources in

South Australia, Australian Geothermal Energy Group Tied Grant.
Executive Summary
A feasibility study has been conducted to assess the benefits of using underground
cooling for the condensers of power plants in regions where environmental constraints
prevent the use of water for cooling. This scenario is expected to apply to many
systems proposed to extract geothermal energy from hot fractured rocks within arid
areas in Australia. The conventional approach for cooling under such circumstances is
by air cooled radiators. However, in these locations typically, the peak in electricity
demand corresponds approximately to the peak in ambient air temperature, which can
be about 45C. This leads to reduced output from the power cycle when electricity
prices are highest. In addition, these radiators require high capacity air fans which can
consume up to 10% of the energy generated.
The concept of underground cooling is to use a closed loop system that utilises the
earths soil for thermal storage during the day and releases it to the atmosphere during
the night. This concept relies on a network of pipes to circulate the cooling fluid (e.g.
water) and transfer its heat to the surrounding soil. During the night, the air
temperature drops and the hot soil transfers its heat to the surrounding air. Since the
soil has a relatively low heat capacity (potential to store energy) the length of the
pipes in such a network can be quite long and be spread over a large area. However,
neither of these two drawbacks was found to limit the feasibility of the concept.
Data were collected of the soil properties and temperature at different depths for
typical sites in South Australia. Weather data were also obtained on an hourly basis
for a typical year. These data were used to calculate the amount of energy gained and
lost from an underground system for a 5MW plant in the Cooper Basin area. Finite
Element Analysis software was used in this study since it can provide an accurate and
time dependent prediction of the rate of heat transfer and temperatures of the cooling
fluid and the soil. These runs were performed for long periods to achieve the required
steady state conditions and account for the worse case scenario in the hottest
conditions of the year.
The case study considered above was used to determine the feasibility of such an
approach and not to provide an optimized design. Hence some simplifications to the
model were made and worse case scenario was considered in terms of weather,
radiation flux and thermal cycle operation. It was estimated that the depth required to
achieve the largest temperature loss from a long 50mm pipe is ~0.1m and that the
overall length of pipe required for this power plant is approximately 25km. It was also
found that a distance between the pipes should be at least 4-5 diameters to minimize
the interference between them. In such a system we will require an area of ~5000 m2.
An indicative cost analysis was performed for this underground cooling system and a
comparison was made with an equivalent air cooled system. The costed system is by
no mean optimized and these figures should be only used as a rough guide. It was
estimated that the capital costs for this underground cooling system is $130,000,
which is significantly less than the estimated $314,000 for the air cooling system. It
was also estimated that the yearly operational cost for the underground cooling system
would be $100,000 as compared with $179k for the air cooled coils. Similarly, the
maintenance cost is anticipated to be lower for an underground system than for an air
i
cooled system. The impact of the condenser temperature on power station output has
not been calculated for the two systems. Nevertheless, the underground system will
maintain a more constant output and will provide greater output during maximum
day-time temperature, since it provided 20C lower temperature at these conditions.
In summary, the study predicts that the concept of underground cooling is likely to be
technically and economically feasible for the conditions chosen for this study.
Nevertheless, as with all models, a number of simplifications and approximations
have been required, so that further model development and experimental validation
are required to help better estimate the benefits. Further work is also required to
optimise the design and further increase benefits of this system with a view to
justifying a demonstration prototype.
ii
Acknowledgements
The authors acknowledge the financial support of the Primary Industry and Resources
in South Australia, PIRSA under the Australian Geothermal Energy Group Tied Grant
scheme. The Project title: "Preliminary assessment of the potential for underground
cooling on power cycle design" and the project ID is TIG 6.2. Most of the research in
this report was conducted by Sean Watson and Andrew Heath as part of their Honours
Project at the School of Mechanical Engineering.
iii
Contents
1 INTRODUCTION.................................................................................................................. 1
1.1 BACKGROUND AND SIGNIFICANCE ...................................................................... 1
1.2 AIMS OF PROJECT....................................................................................................... 2
2 BACKGROUND LITERATURE .......................................................................................... 3
2.1
HOT FRACTURED ROCKS .................................................................................. 3
2.2
SOIL HEAT STORAGE.......................................................................................... 3
2.2.1 IDENTIFYING SOIL PROPERTIES ..................................................................... 4
2.2.2 HEAT BALANCE................................................................................................... 4
2.2.3 COMPARISON OF EXPERIMENTAL & MEASURED RESULTS .................... 5
2.3 ESTIMATING SOIL TEMPERATURE ........................................................................ 5
3
METHODS....................................................................................................................... 7
3.1
MODEL ................................................................................................................... 7
3.2
1D LINE MODEL ................................................................................................... 9
3.3
STEADY STATE 1D VALIDATION MODEL.................................................... 10
3.3.1 SIMPLIFIED MODEL .......................................................................................... 10
3.3.2 FINITE ELEMENT MODEL................................................................................ 12
3.4
TRANSIENT 1D LINE MODEL .......................................................................... 13
3.5
2D LATERAL MODEL ........................................................................................ 14
3.5.1 CONSTRUCTION OF THE FINITE ELEMENT MODEL ................................. 14
3.6
2D LONGITUDINAL MODEL ............................................................................ 17
3.6.1 CONSTRUCTION OF THE FINITE ELEMENT MODEL ................................. 17
3.7
COST ANALYSIS................................................................................................. 21
4 RESULTS AND DISCUSSION........................................................................................... 22
4.1
1D TRANSIENT MODEL RESULTS .................................................................. 22
4.2
DISCUSSION: 1D TRANSIENT MODEL........................................................... 25
4.3
2D TRANSIENT LATERAL MODEL RESULTS.............................................. 26
4.4
2D TRANSIENT LATERAL DISCUSSION ........................................................ 30
4.5
2D TRANSIENT LONGITUDINAL RESULTS .................................................. 32
4.6
2D TRANSIENT LONGITUDINAL DISCUSSION ............................................ 34
4.7
RESULTS OF COST ANALYSIS ........................................................................ 35
4.7.1 COST ESTIMATE OF UNDERGROUND HEAT EXCHANGER ..................... 36
4.7.2 COST ESTIMATE OF AIR-COOLED HEAT EXCHANGER............................ 37
4.7.3 COMPARISON OF COSTS & ANALYSIS......................................................... 38
4.8
EXAMINATION OF FEASIBILITY .................................................................... 39
5
SUMMARY AND CONCLUSIONS ............................................................................. 41
6
FUTURE WORK............................................................................................................ 42
7
REFERENCES ............................................................................................................... 43
APPENDIX A ......................................................................................................................... 45
HEAT CONVECTION COEFFICIENT CALCULATION........................................... 45
APPENDIX B.......................................................................................................................... 46
DATA USED IN FEA TRANSIENT MODELS................................................................ 46
APPENDIX C.......................................................................................................................... 48
COST ANALYSIS QUOTES ............................................................................................. 48
iv
LIST OF SYMBOLS
Symbol
Description
Units
Enthalpy, convection coefficient
kJ/kg, W/m2K
Depth of pipe underground, length of pipe
Qcond
heat flux due to conduction
Qrad
Heat flux due to radiation
Qconv
Heat flux due to convection
Qemit
Heat flux emitted from the soil
Tamb
Ambient air temperature
K or C
Ti
Inlet water temperature
K or C
To
Outlet water temperature
K or C
Tpipe
Pipe temperature
K or C
Tsoil
Soil temperature
K or C
absorptivity
Emittance
Thermodynamic efficiency
Thermal conductivity
W/mK
Density, reflectivity
kg/m3, %
Stefan-Boltzman constant
Wm-2K-4
Cp
Specific heat
J/kg.K
Diameter of the pipe
m&
Mass flow rate
kg/s
Ha
Absorbed solar energy
W/m2
Nu
Nusselt Number
Pr
Prandlt Number
Re
Reynolds Number
Work
kW
1 INTRODUCTION
1.1 BACKGROUND AND SIGNIFICANCE
The growing recognition of the adverse environmental impacts of traditional power
generation techniques is driving the development of alternative, carbon-neutral
methods of power generation. One such approach is to exploit geothermal energy, and
of particular interest in Australia, is a method of power extraction known as Hot
Fractured Rock (HFR) technology. HFR extracts heat that can be found several
kilometres below the Earths crust. Water is forced at high pressure through a series
of boreholes down into the hot rock, where it is heated and re-circulated to the surface
via adjacent borehole(s). Electricity can then be generated from one of several thermal
power cycles by a turbine.
While the HFR technology is of interest throughout Australia, it is particularly well
suited to the Cooper Basin in South Australia because of the relatively shallow depth
of the hot rocks there compared with many other HFR sites. However, utilising the
HFR technology in this region poses numerous technical challenges, not the least of
which is the method by which the working fluid in the power cycle will be cooled.
The need for cooling in any thermal power cycle has driven many conventional power
plants to be located close to a river, lake, or ocean to provide an environmental heat
sink (Department of Environment, 2001). Where this is not realistic, cooling towers
are almost invariably used to provide greater cooling than is possible by air-cooling
alone, utilising the evaporation of water, much like a large evaporative air cooler.
However, the water consumption due to evaporation and fouling losses, even for a
conventional power station with comparatively high efficiency, is typically 1363
L/MW.h per day (Ricketts et al., 2006). Indeed, the recent increase in demand for aircooled condensers is a direct result of water shortages at potential plant sites and
increasing government legislation limiting the use of water in the wet cooling systems
(Ricketts et al., 2006). Unfortunately, the regions in which geothermal heat is the
most viable are arid or semi-arid, with no readily available source of surface water for
cooling. While underground water is available, notably from the Great Artesian Basin,
it is unlikely that environmental regulators will allow it to be utilised because of the
very large water consumption required by any large thermal power plant. Therefore,
conventional water-based cooling methods are not expected to be possible for these
power plants.
A possible alternative method of heat dissipation is through the use of conventional
fin-and-tube air cooled heat exchangers, in which air is passed over a heat exchanger
to remove heat from the working fluid. Fins on the surface of the heat exchanger are
typically used to improve the heat transfer by increasing convection and radiation
away from the surface. Such systems are used in a number of geothermal plants, e.g.
in the Mokai plant in New Zealand. However, in the areas in which HFR is expected
to operate, daily ambient air temperatures can reach 45C in the summer, in which
case the minimum temperature of the working fluid temperature would typically reach
some 48-50C, owing to the fact that the effectiveness of a heat exchanger is always
less than 100%. Such high condenser temperatures result in a low efficiency of the
power cycle. For example, Langman et al (2008) estimate that the output from a
1
geothermal plant at a typical site in South Australia would drop by 40% as the
ambient air temperature is increased from 15 to 45 C. Furthermore, the peak demand,
and hence peak prices, in the national grid are greatest during the very time when the
output is lowest. Such a scenario could significantly influence the economic viability
of a plant. Fans may be employed to force air over the fins to improve the
performance of the heat exchanger, but consume large amounts of energy and hence
reduce the net amount of electricity produced by the plant.
A potential solution to the problem of cooling the working fluid is to install a heat
exchanger underground where temperatures are lower more stable. Heat in the
working fluid may then be rejected to the soil and in turn dissipated to the atmosphere.
For this reason, the aim of the present investigation is to assess the feasibility of
underground cooling for such a geothermal plant.
1.2 AIMS OF PROJECT

The aims of this project as listed in the proposal are:
1. To undertake a review of the types of underground cooling systems presently
available. This literature review will aim to identify:
a. Methods previously used to enhance the heat transfer coefficient to the
earth
b. The type and properties of soils from potential sites for Geothermal
exploration
c. The relative performance of alternative filling materials in terms of
heat transfer coefficient
d. Comparative performance of different network configurations, and
their influence on the thermal gradients in the earth.
e. Relative trends in excavation costs associated with different network
designs
2. To undertake preliminary designs of an underground condenser and an aircooled condenser of the same capacity.
3. To undertake a preliminary and indicative relative cost estimate based on
nominal and/or typical conditions.
4. To identify opportunities for future cost-savings through further research.
2 BACKGROUND LITERATURE
2.1
HOT FRACTURED ROCKS
Hot Fractured Rocks (HFR) technology is a recent development in geothermal power

generation that utilises the heat generated below the earths crust. If used in
conjunction with a binary power cycle, water will be fed down into the crust, deep
below the surface (approximately 4-6kms) of the earth, where the rocks have a
temperature greater than 270C, as shown in Figure 1. The rocks have special
radiogenic materials that produce their own heat, which means they can be utilised to
heat up objects and fluids around them (Geodynamics, 2008). This water will then be
held in a thermal reservoir where it will be heated to between 200C and 250C (The
efficiency of a thermal power cycle increases with the exit temperature of the
geoliquid, but the temperature gradients through the rock are non-uniform, resulting
in an optimal depth at which heat can economically be extracted). The water will then
be extracted and pumped up to the surface and used to provide the heat source for the
power cycle. This heat source requires no fossil fuel and so does not produce any
combustion-generated emissions. Nevertheless, emissions will arise from the
manufacture and construction of the site and the drilling, etc. In addition, the water
consumption is expected to be high, with typically 5-10% of the geoliquid flow-rate
anticipated to be lost through the porous underground heat exchanger. Depending on
the design, HFR systems can also result in other environmental impacts. Flash cycles,
for example, may extract some non-condensable gases, such as CO2 and CH4.
Likewise, some plants employ acid injection to prevent fouling from dissolved silica
onto heat exchanger equipments (De Pippo, 2005).
Figure 2.1 Schematic diagram of a HFR geothermal power plant [Geodynamics, 2008]
2.2
SOIL HEAT STORAGE
The heat exchange between the earth and the atmosphere depends on the net balance
of radiant and convection processes, which vary with the time of day and the season.
3
In addition, the large thermal mass, or heat capacity, of the earth results in a time lag
between the temperature of the earth and that of the ambient air. For these reasons the
use of soil to reject the heat from the condenser is known as a thermal storage cycle.
Typically, there is a net storage of heat in the soil during the day and a net dissipation
of heat from the soil to the atmosphere at night.
There are many factors affecting the capacity of the soil to store heat. These factors
include water content, phase composition, porosity, convection, conduction, heat
capacity and temperature variation, daily and seasonally (Marra, 2006). The main
factors, however, are specific heat, density, thermal conductivity, solar radiation
intensity, wind speed, rain, humidity and ambient temperature (Nassar et all, 2005).
2.2.1 IDENTIFYING SOIL PROPERTIES

The performance of an underground heat exchanger does not depend only on the
temperature of the soil at different depths. It also depends on specific heat, density
and thermal conductivity, which are soil specific properties. These parameters are
therefore needed to determine the thermal storage capacity of the soil. Soil comprises
of many elements such as sand, air and water (Nassar et all, 2005). The volumetric
percent of each element can be calculated (Nassar et all, 2005) as follows:
( CP ) = f s ( CP ) s + f a ( C P ) a + f w ( C P ) w
(2.1)
where f is the volumetric percent and the subscripts s, a and w are sand, air and water
respectively. The sum of these respective volumetric percentages is equal to unity. i.e.
fs + fa + fw = 1
(2.2)
2.2.2 HEAT BALANCE

The temperature of the soil at a given depth and time can be determined theoretically,
given suitable data, using the unsteady heat equation (Nassar et all, 2005) as follows:
dT
(2.3)
dt
where Ha is the absorbed solar energy at the surface of the soil, qs" is the heat flux due
H a q s" q L" q v" = ( C P )x
to conduction that is transferred to the soil, q L" is the heat flux lost due to radiation
and convection from the soil surface to the surrounding air, q v" is the heat flux lost
due to evaporation, ( C P ) is the heat content of the soil components as given by
Equation 2.1, x is the depth below the surface of the soil and dT/dt is the soil
temperature change with time. Nassar et al. 2005, have found correlations between
soil depth and density, thermal conductivity and specific heat that satisfy Equation 2.3
for the conditions that apply in Tripoli City, Libya. These values have been graphed
from the corresponding formulas and are shown in Figure 2.2.
Figure 2.2: Soil thermal properties vs depth for soil in Tripoli City, Libya[Nassar et al, 2005]
Since the data in Figure 2.2 is site specific, it cannot be directly translated to Australia.
Nevertheless, the trends are expected to be similar in most locations.
2.2.3 COMPARISON OF EXPERIMENTAL & MEASURED RESULTS

Equation 2.3 was validated experimentally by Nassar et al (2005), using a borehole of
4m depth and 2m width, instrumented with four arrays of thermocouples in each
corner at 0.5m intervals. Data was recorded hourly. The borehole was dug in Tripoli
City, Libya. Figure 2.3 displays the temperature distribution over an entire year for
differing depths. It is evident that the temperature distribution exhibits only seasonal
variation, with little daily variation, at depths greater than 1.5m. These curves can be
used to verify the above calculated values of soil thermal properties.
2.3 ESTIMATING SOIL TEMPERATURE

Estimating temperature data is very difficult without time-consuming, labourintensive field tests and expensive experiments. Perry, Ristaino and Wu (1995)
suggest a method to calculate the approximate soil temperature without conducting
the field tests and experiments. Their method requires only hourly measurements of
ambient air temperature, global radiation, dew point, wind speed and rainfall. This
model was found to work well except where there is rapid change in air temperature
and global radiation. However, even in the worst case, the error was at most 11%.
The general equation to estimate soil temperature is as follows:
T ( Tt )
C
=
t
Z
(2.4)
where C is the volumetric heat capacity of the soil, T is the soil temperature, is the
differential operator, t is time, z is depth and is the thermal conductivity.
Figure 2.3: Temperature distribution for varying depths [Nassar et al, 2005]
Perry, Ristaino and Wu (1995) have developed a detailed approach to calculate

temperature from this equation, and their results are compared with measured data at a
depth of 10cm below the surface in Figure 2.4.
Figure 2.4: Comparison of estimated versus measured temperature [Perry, Ristano & Wu, 1995]
It can be seen from Figure 2.4 that the estimated values agree well with the measured
values. Although the peaks and troughs differ by 1 or 2 C, they occur at
approximately the same time of the day. This implies that the temperature calculated
by this approach is accurate to within 2C.
METHODS
The heat transfer processes within an underground heat exchanger are time dependent
and three dimensional. Many analytical solutions are not able to account for such
variations. Here we have selected a modelling strategy that accounts for temperature
variation and heat flow over 24 hours, or over a period of many days. The problem
can be described in a simplified way by assuming that:
the soil is homogeneous;
there are no phase changes in water, or latent heat effects;
the soil emissivity, absorptivity and reflectivity are constant;
convection over the surface is a function of wind speed;
the pipe depth is fixed, and
the effects of the pipe wall on conduction are negligible.
3.1
MODEL
A block of soil with sides of approximately equal length, shown in Figure 3.1, is
analysed. A pipe representing the heat exchanger is buried underground at a certain
depth, l, and the working fluid of the power cycle is passed through the pipe at a
temperature Ti. Since the soil is at a lower temperature than the pipe, heat energy is
lost to the soil, such that the fluid exits the pipe at some lower temperature To. It
should be noted that the boundaries of the block of soil are assumed to be semiinfinite, and thus heat may be lost through the faces of the block with negligible
resistance.
Figure 3.1 Schematic diagram of a pipe buried in soil
The pipe depth l is an important parameter and will have a crucial effect on the
behaviour of the model. It is known that fluctuations in temperature over a period of
time decrease with soil depth, depending on the properties of the soil (Pavelka et al.
2006). The measured change in temperature over the period of one day for various
soil depths is presented in Figure 3.2. Thus, at a depth of just 30cm the temperature
fluctuation for soil over a day can be small. However, these data were obtained for
soil properties emulating that of a forest bed and at a different geographical location,
hence climate, than those being considered for this study. Therefore, for initial
modelling purposes l will be set at 1m to provide reduced effects of surface
temperature fluctuation on the model. Recorded weather data are also used in the
model.
Figure 3.2 Temperature fluctuations at various soil depths over 24 hours period [Paveleka et al.,
2006
For the present comparative assessment, it is sufficient to assume a typical

temperature for the working fluid entering the condenser. We have chosen this to be
100 C. This is typical of the temperature at which the geoliquid is expected to be
returned underground from EGS systems under conditions in central Australia
(Langman et al, 2008). While being somewhat higher than expected condenser
temperatures, for the present comparative purposes it is sufficient to ensure that both
the geothermal and air-cooled temperatures are based on the same reference condition.
The heat being lost to the soil through conduction is termed Qcond in Figure 3.1 and
depends on a number of parameters, including the conductivity of the soil. The
conductivity of soil changes as the moisture content of the soil increases or decreases
and is largest when the soil is wet. However, as rain is rare in the arid regions being
considered, the soil conductivity is assumed to be that of dry soil, and is assumed to
be constant. Note that this assumption is conservative, since the presence of moisture
will increase the thermal conductivity. One report suggests that thermal conductivity
values for sandy loams range from 0.54W/mK to 1.94W/mK (Abu-Hamdeh and
Reeder, 2000). Mills (1999) suggests that for dry soil a value of 1.0 W/mK is
appropriate, although without providing details of the type of soil for which this value
was obtained. Exact data for the geomorphology of the Cooper Basin region is
difficult to obtain, as reports suggest that the region contains both wetlands and desert,
which have widely varying soil properties (Burdon, 2006). To allow for a
conservative solution for conduction then, it was decided to use a value of 0.75W/mK
for the thermal conductivity of all models, which is typical of sandy loam soils.
The heat transferred to the surface by radiation from the Sun during the day is termed
Qrad, shown in Figure 3.1. Solar radiation is made up of two components, namely
8
direct radiation and indirect radiation. For the purposes of the present model, the
direct and indirect radiation are grouped into a single quantity. Data for the daily and
seasonal variation in Qrad, are readily available.
The infra-red radiation from the soil is shown in Figure 3.1 as Qemit. It is assumed that
the soil is a gray body and will emit radiation as a function of the emittance of the soil
and the temperature difference between the soil and the air. The emittance of the soil
is assumed to be constant and equal to 0.75 (Mills, 1999). Data for the daily and
seasonal variation in ambient temperature are readily available.
Heat is also dissipated through convection from the surface, termed Qconv in Figure
3.1. This depends on the wind speed of the air flowing over the surface and the
ambient air temperature. The average daily wind speeds for the month of January
were obtained from Energy Efficiency and Renewable Energy (2008) and were
averaged to provide a monthly average value. This monthly average value was used to
calculate the convective heat transfer coefficient. Details of this calculation are shown
in Appendix A.
To further examine the effect of these variables on different aspects of the model, it is
necessary to divide the problem into smaller models. Considering the block of soil in
Figure 3.1, cross-section cuts can be made through the block to obtain two separate
models. These models include the lateral and the longitudinal cross sections and they
will be discussed in further detail below.
3.2
1D LINE MODEL
In investigating a complex problem it is good practice to initially reduce the problem

to the simplest form possible. For this study, a 1D line model was chosen as this
embodiment. Here a cross section of the pipe face and surrounding soil is obtained. If
this section is viewed from the side, then Section C-C becomes apparent (Fig. 3.3).
(a) Section A-A
(b) Section C-C
Figure 3.3 Schematic diagram of the 1-D model
In Figure 3.3(b), Section C-C is a 1D line model that extends from the top of the pipe
(1) to an arbitrary point within the surrounding atmosphere (3). For modelling
purposes, the soil depth is set to 1m and the distance above the surface (3) is set to
0.5m. An important feature to note is that surface node (2) is modelled slightly below
the surface, rather than on the soil surface. Hence it only receives a portion of the
incoming radiation, dependent on the absorptivity of the soil. This, in turn, depends
on the location and moisture content of the soil. Hays et al. (2001) indicated that
sandy type soil has a reflectivity of approximately 30%. As absorptivity, , is
defined as being equal to (1- ), then it can be assumed that approximately 70% of the
global radiation is absorbed for this node for the rest of this report.
3.3
STEADY STATE 1D VALIDATION MODEL
3.3.1 SIMPLIFIED MODEL

The use of FEA software can provide powerful modelling capability, provided it is
validated. For this reason, a simple steady state problem was constructed and
compared to simplified models for verification purposes. Figure 3.4 presents a
diagram of the validation problem.
10
Figure 3.4 Schematic of verification model
This problem considers a steady state analysis of a buried pipe that is subjected to
known boundary conditions. (Refer to Section 2 for an explanation of
terminology).The goal of this validation problem is to determine the temperature of
the surface, T, and to compare it to the results from FEA.
The heat convection coefficient h used was calculated using equations for forced
convection over a flat plate, and assumes a wind speed of 10m/s. The wind speed of
10m/s is equivalent to a strong breeze. Although arbitrary, this is sufficient for
validation. This value is found to be 37 W/m2K as shown in Appendix A.
The known thermal properties and other required data for this problem are shown in
Table 3.1 and Table 3.2:
Table 3.1: Soil and air thermal properties

Symbol
Soil
Air
k(W/mK)
0.75
0.0271
C(J/kgK)
1000
1900
1500
1.225
(kg/m3)
0.77
Table 3.2: Temperature data and other parameters

Term
Value
Tamb
313K
Tsoil
298K
Tpipe
373K
Asoil
1m2
hair
36.99 W/m2K
5.67E-08 W/m2K4
In this problem, it is assumed that all properties are constant and that a steady state
analysis applies. For radiation it is assumed that the air to ground shape factor is
11
approximately unity and does not need to be included. Also, for ease of calculation a
unit cross sectional area is assumed.
The surface temperature, T, is required. Equations 3.1 to 3.4 were obtained from Mills
(1999). An energy balance on the surface of the soil produces:
qcond qrad - qconv = 0
(3.1)
where q rad is the radiation incident on and leaving the soil, and is defined by:
q rad = E b1 s G s G
(3.2)
and qconv is the convective heat loss on the soil, and is defined by:
q conv = hT
(3.3)
and q cond is the conductive heat gain through the soil, and is defined by:
q cond = k
T
x
(3.4)
Expanding Equation 3.1, leads to:

T
+ hair A(Ts Tamb ) + E b1 s G s G = 0
(3.5)
x
For this problem, the combined global radiation term s G s + G is assumed to be
equal to 870W/m2. This corresponds approximately to the peak solar loading on the
soil for a given day. The parameter Eb1 is the grey body emissive power of the soil.
Hence one can find iteratively that Ts is:
kA
Ts = 334K
(3.6)
This answer is compared to the results obtained from FEA in the following section.
3.3.2 FINITE ELEMENT MODEL

To build the ANSYS model appropriate for this problem it was necessary to divide
the problem into parts. Each of the terms for radiation, conduction and convection are
calculated separately using the thermal element types LINK31, LINK32 and LINK34,
respectively, within ANSYS. In order to use these element types, nodes were created
as shown in Figure 3.3. The air node presented in that figure was split into two
separate nodes to allow for the differing effects of convection and radiation. The
nodes were then linked by line elements and the appropriate element types were
applied. To summarise, the 1D ANSYS line model consists of four nodes representing
the top of the pipe, the surface of the soil, and two nodes representing the air, which
are linked by three elements to allow for conduction through the soil and convection
and radiation to the surrounds. This geometry is presented in Figure 3.5.
12
Figure 3.5: 1D ANSYS line model geometry
The material properties shown in Table 1 and boundary conditions shown in Table 2
were then applied to the model. Solutions were obtained for the temperature at all
nodes and are shown in Figure 3.5:
Table 3.3: ANSYS obtained nodal temperatures for the 1D steady state case
Node
1
2
3
4
Temperature (K)
373
334
313
313
The temperature of the surface, Node 2, was calculated by this ANSYS model to be:
Ts = 334K
This answer corresponds to the solution calculated above in Equation 3.6, thus
validating the FEA model. Further analysis using the FEA software may now be made
with confidence.
3.4
TRANSIENT 1D LINE MODEL
The 1D transient model is similar to the steady state model examined above. That is,
the transient model has the same geometry and types of boundary conditions as the
steady state model but is solved as a function of time. The transient model accounts
for variations in the ambient temperature and global radiation. To this end, data for
the hourly average ambient temperature for all days in the month of January was
13
obtained (EERE, 2008), and is shown in Appendix B. This temperature data was
obtained from a weather station in the town of Oodnadatta, which is representative of
summer daily temperatures in the Far North East of South Australia. The choice of
January was due to it being in the middle of summer with the hottest daily
temperatures. This corresponds to the worst loading conditions for the heat exchanger.
Global radiation data was obtained for the model (EERE, 2008) as an average hourly
radiation for all days over the month of January (see Appendix B). This data was also
acquired from an Oodnadatta weather station for the same reasons as mentioned
above.
The daily average wind speed for the month of January was used to calculate the heat
convection coefficient to be 23.6W/m2K.
These data sets were then applied to the appropriate nodes as tabular loads, and the
boundary conditions used in the steady state model were also applied. The model was
then solved using a transient solver for time periods of 1 day, 10 days and 30 days.
The results are presented in the Results section.
3.5
2D LATERAL MODEL
The problem examined next was a lateral cross section of the soil block, shown in
Figure 3.6 as Section A-A. This model considers an infinitely thin section that shows
the pipe cross section to a front-on observer. This model is used to examine the
temperature profile of the soil surrounding the pipe over time, assess whether the
calculations reach a steady state and the effect of pipe spacing.
Figure 3.6: Section A-A: A lateral cross section of the pipe buried in soil
3.5.1 CONSTRUCTION OF THE FINITE ELEMENT MODEL

Owing to constraints of the ANSYS package, to model an infinite boundary condition,
it is necessary for the boundary of the problem to be circular, as shown in Figure 3.7,
rather than square, as modelled above.
14
Figure 3.7: Reconfigured lateral cross section of the pipe buried in soil
Shown in the reconfigured geometry is a pipe of diameter d buried in soil at a depth l

surrounded by a circular boundary with radius r. The pipe diameter was arbitrarily set
to 0.2m. The effect of varying the pipe diameter d on the results will be examined in
future work. The soil depth l was set to the same value used for the 1D line model of
1m, and the radius of the infinite boundary was entered as 2m. The geometry was then
meshed using triangular elements, although the shape of the mesh elements used in
this case is not important to the final solution. This is because the heat transfer process
in the soil is that of conduction, which is a linear process. Therefore, as long as the
elements are not highly skewed, the solution would be expected to be the same within
the soil for triangular and quadrilateral elements. The mesh employed in the model
can be seen in Figure 3.8. With the model now able to properly allow for infinite
boundary conditions, the semi-circular edge was selected and re-meshed with INFIN9
elements.
Figure 3.8: Lateral geometry with applied mesh
15
As with the 1D line model, the three thermal loads of conduction, convection and
radiation will act on the model. Conduction is provided for by using the ANSYS
element type PLANE55, which was applied to all elements in the soil. It was assumed
that the cross sectional area for conduction was 1 m2, so that results would also be in
terms of a unit cross sectional area (for example, W/ m2 would be the output units for
heat flux on the soil).
Convection was incorporated using a slightly different method. To compute
convective loads for a 2D model ANSYS employs an element type called SURF151.
To use SURF151, the elements on the surface for which convection acts are first
selected. An extra node, created some distance away from the surface, is then selected
and the surface elements re-meshed. The extra node is used for measuring the
temperature gradient between the surface and the surrounds. Therefore, it represents
an arbitrary point in the ambient air with temperature Tamb. A convective load with an
appropriate heat convection coefficient is then applied to the SURF151 elements on
the surface. Convection may now be calculated for the model.
Radiation is accounted for in a similar way to convection. SURF151 is also used in
this case, but options are selected in ANSYS to specify radiation loads rather than
convection loads. The surface line is again re-meshed with SURF151 elements in the
manner described above, with a different node selected as the extra node to which the
temperature gradient will be computed. A specific radiation load does not need to be
applied to the surface, as radiation is calculated automatically from the material
properties and other boundary conditions entered. The extra nodes used for radiation
and convection can be seen in Figure 3.8 the surface of the soil.
Boundary conditions similar to the 1D transient line model were then applied to the
model. The time varying global radiation load was applied to the surface of the soil,
again assuming that the surface in fact only receives around 70% of this load due to
absorptivity effects. Hourly temperatures were applied to the extra nodes referred to
above, and the soil temperature was initially set to 25C. The soil material properties
presented in Table 3.1 were also entered into the model.
The transient solution option was then turned on, and the model way solved for timesteps of 1 day, 10 days, and 30 days. The solutions generated are presented in the
Results section.
16
3.6
2D LONGITUDINAL MODEL
The final model to be examined was a longitudinal cross section cut of the buried pipe.
As can be seen in Figure 3.9, this model represents a pipe buried beneath the soil at a
certain depth for a length of 50m. This model was used to determine the overall
temperature drop over a 50m length of pipe and hence, the optimal depth of the pipe
beneath the soil was evaluated. The optimal pipe depth was calculated by considering
the overall temperature drop at varying depths and choosing the depth at which a
maximum temperature drop was achieved. The longitudinal model was also used to
determine the overall length of pipe needed to adequately cool the working fluid to
the desired temperature. This was achieved by obtaining the exit temperature of the
working fluid in the pipe for successive 50m sections of pipe. A relationship between
temperature loss and pipe length could then be derived, and the overall length of pipe
determined. These concepts will be developed and explained further section 4.5.
Atmosphere
Depth
Pipe
Soil
50m Long
Figure 3.9: Section B-B: A Longitudinal Cross Section of a pipe buried in the soil
3.6.1 CONSTRUCTION OF THE FINITE ELEMENT MODEL

The 2D longitudinal model was constructed in a similar way to the 2D lateral model,
but instead had to incorporate the effect of the working fluid flowing through the pipe.
The effect of the fluid flow on the model will be discussed in more detail later. Firstly,
it should be noted that as before, this model used the infinite boundary element to
allow heat flow through boundaries and not be reflected back. In the same way, this
then required the infinite boundary to be of a hemispherical shape. Figure 3.11
displays the reconfigured geometry of the longitudinal model.
In Figure 3.10, the pipe may be observed as a section along the length of the model, l,
at a depth h, and with diameter d. The diameter of the radial boundary, r, was set to be
the same as the length of the pipe, l, for purposes of symmetry. The length of the pipe
was set to be 50m, which was chosen arbitrarily to represent a section of pipe. For
17
reasons of cost, in this model the pipe diameter was selected to be 50mm. This will be
discussed in further detail in the Section 3.7.
Figure 3.10: Reconfigured longitudinal cross section of the pipe buried in soil
The soil was meshed using quadrilateral elements for this model, as this produced a
more even mesh for the given geometry. As stated in Section 3.5.1 however, it is
important to note that the same solution may be obtained with triangular or
quadrilateral elements, provided the elements are not highly skewed. The mesh
employed in the model can be seen in Figure 3.11. The radial edge of the geometry
was re-meshed with INFIN9 infinite elements, as used in the lateral model.
The three modes of heat transfer are again required to be implemented for this
particular model. As before, conduction was provided for by using the ANSYS
element type PLANE55, which was applied to all elements in the soil. For the same
reasons as given above, it was assumed that the cross sectional area for conduction
was 1 m on the surface of the soil to give answers in terms of unit area.
Convection and radiation were employed for the longitudinal model in precisely the
same way as for the lateral model. Note again that the extra nodes used are
representative of arbitrary points in the ambient atmosphere above the soil, and have
time-varying daily temperatures applied to them.
18
Figure 3.11: Finite element model of the Longitudinal Cross Section
The boundary conditions applied to the model were similar to that of the lateral model,
but have some important differences. The same time varying global radiation load
was applied to the surface of the soil, and the soil temperature itself was assumed to
be initially set to 25C. The soil material properties presented in Table 3.1 were also
used. Recalling that the parameter of interest for this model is the exit temperature
from the pipe, the only boundary conditions applied to the pipe itself were the
temperature at the inlet and the mass flow rate of the water. As with the lateral model,
the temperature at the inlet was assumed to be 100C. The flow rate required was
calculated from the formula for steady-flow heat transfer for incompressible liquids,
as given by Mills (1999).
The formula is stated:
Q& = m& c p T
(3.7)
where is the heat transferred to or from the fluid [kW], is the mass flow rate [kg/s], is
the specific heat of water [J/kgK] and is the temperature difference [K] of the fluid in
transferring . Rearranging Equation 3.7 for m& , it can be obtained that:
m& = Q& / c p T
(3.6)
If, as stated previously, the system is required to reject 5MW of heat, and it is
required to cool the water from 100C to 30C, and the specific heat of water is
19
assumed to be approximately 4200 J/Kg K, then the mass flow rate of the water into
the system may be given as:
m& = (5 10 6 ) /( 4200 (100 30)) 16kg / s (3.7)
This mass flow rate was entered as one of the boundary conditions associated with the
longitudinal model.
Also required in the finite element model is the convective heat transfer process from
the working fluid to the pipe. The working fluid in the pipe is represented as a series
of nodes connected by FLUID116 elements, which account for fluid mass transport
effects. The FLUID116 elements are also connected to adjacent nodes representing
the pipe/soil boundary. This may be observed in Figure 3.12.
Figure 3.12: Close View of Nodes and Elements
To compute the convective heat transfer process, ANSYS requires a mathematical

relationship to be defined between the nodes of the pipe fluid to the nodes at the
pipe/soil boundary. This allows the ANSYS program to calculate the heat convection
coefficient for the properties of the working fluid as they change with temperature.
The relationship chosen for this process is the Dittus-Boelter correlation for fully
turbulent flow in a smooth pipe (Crawford, et al., 2004), and is stated
Nu D = 0.023 Re 0D.8 Pr 0.4
(3.8)
Fully turbulent flow in a pipe may be characterised as turbulent if the Reynolds

number is above 2300. For flow at 16 kg/s in a 50mm diameter pipe at approximately
100C, and with a density of 958 kg/m3, the velocity of the flow may be given by
20
v=
(m& / )
= 8.51 m / s
2
d
4
(3.9)
Entering in 3.12 and other relevant parameters into the equation for Reynolds number
yields
Re D =
vd
1.4 10 6 > 2300
(3.10)
From Equation 3.10 it may then be assumed that the flow in the pipe is fully turbulent
and that the relationship defined in Equation 3.8 is applicable.
The solutions generated are presented in the Results section.
3.7
COST ANALYSIS
To perform a cost analysis, industry experts with relevant experience in cost

estimation have been consulted, and quotes obtained. The aspects of the project that
required cost estimates were deemed to be the capital cost of the underground heat
exchanger, and the on-going costs of running this system. Note that when obtaining
the estimate for the underground heat exchanger, the optimised parameters (such as
pipe length and depth) found by using the methods stated in Section 4.5 were used.
These costs will be compared to costs obtained for an equivalent air-cooled heat
exchanger (including capital and on-going costs).
21
4 RESULTS AND DISCUSSION

Results from the FE analysis are presented in three sub-sections comprising use of the
1D transient model, the 2D lateral model and the 2D longitudinal model. Following
this a cost analysis is presented. These are used to estimate both the technical and
economic feasibility of this approach.
4.1
1D TRANSIENT MODEL RESULTS
The transient model was employed to assess the influence of changes in ambient air
temperature and global radiation through the day, based on hourly averaged data for
the month of January. Simulations were modelled over a single day, 10 days and 30
days. The results for this model are presented in the form of two graphs. The first
graph is the temperature distribution at the surface of the soil over the time period.
The second graph is the heat flow into and out from the surface of the soil. Figures 4.1,
4.2 and 4.3 represent the temperature of the surface over a 1-day, 10 day and 30 day
period respectively.
Figure 4.1: The surface temperature calculated over 1-day by the 1-D transient model over the
50m length
22
Figure 4.2: Surface temperature calculated for 10-days by the 1-D transient model over the 50m
length
Figure 4.3: Surface temperature calculated for 30-days by the 1-D transient model over the 50m
length
23
Figure 4.4: Heat flow through the surface calculated for 1-day by the 1-D transient model over
the 50m length
Figure 4.5: Heat flow through the surface calculated for 10-days by the 1-D transient model over
the 50m length
24
Figure 4.6: Heat flow through the surface calculated for 30-days by the 1-D transient model over
the 50m length
4.2
DISCUSSION: 1D TRANSIENT MODEL
The results from the one-dimensional transient calculations assess the time required
for the system to converge to a steady cycle. When run for only one day (Figures 4.1
and 4.4), the cycle does not converge to this state. Rather, as shown in Figures 4.2 and
4.5, it takes some three days for the system to converge toward a steady state
behaviour. Steady state behaviour is defined as the ability of the model to repeat a
day-night cycle over and over again and produce results that repeat and are equal for
each repeated day. The results also display the type of transient behaviour expected
for the hourly fluctuation in radiation and ambient temperature For example, the
maximum and minimum temperatures (Figure 4.1) correspond to the time at which
there is a change in sign of the heat flux through the surface (Figure 4.4). They also
reveal that the maximum and minimum calculated temperature of the soil is 36C
(309K) and 24C (297K) over the 10-day cycle.
A closer inspection of the 10 day calculations, shown in Figure 4.2 and 4.5, reveals
that the maximum heat flow out of the soil is approximately 400 W/m and the
maximum heat flow into the soil is approximately 820W/m once steady state
behaviour has been achieved. This appears to indicate that more heat is entering the
soil than leaving it, i.e. that more time is required before a true asymptotic state is
reached. This issue is assessed in more detail below.
The 30-day results shown in Figures 4.3 and 4.6 provide some conflicting data with
previous solutions. The maximum temperature calculated over the 30-day period
(Figure 4.3) is 45C (318K), 9C higher than calculated from the 10-day analysis.
Given that the steady state behaviour is achieved within the same time period
(approximately two days), the maximum temperatures for a 30 day analysis should
match that of the 10 day analysis almost exactly. The fact that the temperatures do not
25
match indicates that one of these models is not performing as expected. The minimum
temperature of the 30 day model is also different from that of the 10 day model.
Similar disparity between results for heat flow at the surface of the soil is also
apparent. The maximum heat flow out of the soil for 30-days Figure 4.6 is lower than
that of 10 days Figure 4.5 and the maximum heat flow into the soil is lower than that
of 30-days. These apparent discrepancies might be explained by results presented in
Section 4.5, which finds that steady state is not reached until approximately 18 days.
Apart from these differences, the 30-day results shown in Figures 4.3 and 4.6
correlate well with those from the 1- and 10-day solutions, as expected. The
maximum temperature achieved over 30-days (Figure 4.3) is 318K when steady state
behaviour is achieved. As the model was observed to achieve steady state behaviour
after two to three days, the temperature and heat flux fluctuations are the same as
those presented for the 10-day model.
4.3
2D TRANSIENT LATERAL MODEL RESULTS
The results from the 2D lateral model are important in determining several parameters
of interest. First, results may be obtained for temperature of the surface soil over time,
as well as the temperature contours of the soil itself. The results presented in this
section will also allow for analysis of whether the soil in the model reaches a steady
state condition over time, and the minimum spacing between pipes. Solutions for this
model were generated using the approach explained in Section 3.5 and are presented
below. The first results obtained were temperature solutions for a node on the surface
of the soil that is directly above the pipe. The nodal temperatures were obtained for
durations of 1 day, 10 days and 30 days and are shown on Figures 4.7, 4.8 and 4.9,
respectively.
Figure 4.7: Temperature variation at the surface node as calculated by the 2D lateral model over
1 day
26
Figure 4.8: Temperature variation of the surface node, as calculated by the 2D lateral model over
10 days
Figure 4.9: Temperature variation of the surface node as calculated by the 2D lateral model over
30 days
It was also desired to obtain temperature contours within the soil over time. The time
period used for this analysis was 1 year, which was chosen as being sufficiently long
for the model to converge to an asymptotic cyclical behaviour. Data is presented from
the start, middle and end of this time period in Figures 4.10, 4.11 and 4.12
respectively. Note that an animation of the variations in the contours over this 1 year
period is also available.
27
Figure 4.10: Temperature contours for 2D lateral model at time=900s (<1 hour)
Figure 4.11: Temperature contours for 2D lateral model at time=1.1E+6s (Approx. 13 days)
28
Figure 4.12: Temperature contours for 2D lateral model at time=2.6E+6s (Approx. 30 days)
Figure 4.13: Contour plot corresponding to approximately 1 year
29
It was also deemed necessary to obtain the time history of two nodes directly below
the pipe that are sufficiently far away from the surface as to not be affected by the
solar radiation and convection. The time period used for this analysis was also one
year. The two nodes selected are shown in Figure 4.14.
Figure 4.14: Two nodes selected for analysis
4.4
2D TRANSIENT LATERAL DISCUSSION
The temperatures at a surface node show that the temperature is increasing from the
initial value set at 298K (25C) to a maximum in the middle of the day when the solar
load is at a peak, as expected (Figure 4.7, 4.8 and 4.9). This maximum is
approximately 325K (52C), and is approximately the same for all days, perhaps
indicating that the model is approaching steady state behaviour.
Further examination of the nodal temperature solutions for 1 day, 10 days and 30 days
show a similar behaviour to previous results. It can be seen that the maximum
temperatures for a surface node are increasing for each of the 1-day, 10-day and 30day results, which was also observed for the 1D transient model. As indicated
previously, it is possible that the model has not yet reached steady-state and may have
to be evaluated over a longer time period. It is also noted that the maximum and
minimum temperatures for the 1-day, 10-day and 30-day are the same as the 1D
transient results for the same time period. For example, by comparison of Figure 4.3
and 4.9 it is observed that the maximum temperatures are approximately equal.
Figures 4.10, 4.11 and 4.12 provide insight into how the heat is distributed from the
pipe. From Figure 4.10 it can be seen that heat is localised around the pipe and begins
to spread outwards. Shown in Figure 4.11 is at 13 days into the cycle the soil
temperature some distance from the pipe has increased but still stays relatively cool
when compared to the pipe temperature. This figure also shows that the heat is being
30
transported to the surface of the soil and is being lost to the atmosphere. At this timestep of 13 days, the green contours show that heat is flowing into the surface of the
soil from radiation but does not penetrate the ground to a significant depth. It does,
however, show that the surface of the soil is just as hot as around the pipe, since the
contours have the same colour. This implies that heat will not be lost to the surrounds
when radiation is at such a level of intensity. This means that the pipe will simply heat
up the surrounding soil during the day when radiation from the sun is acting upon the
surface of the soil. Figure 4.12 indicates that, although the system is not showing
steady state behaviour, the area of interest around the pipe is reaching quasi-steady
state behaviour because the contours around the pipe, when comparing the 13 day
contours to the 30 day contours, have roughly the same diameter. The temperature
contours at about 1 year are shown in Figure 4.13. In this figure it is observed that the
temperature in the soil surrounding the pipe has lowered significantly, most likely due
to the fact that the ambient conditions are much more favourable. That is, as the 1year model begins in January and ends in December, this implies that the highest
loading on the model will be just after the beginning of the year, when the ambient
conditions (and hence heat loading) will be most adverse. The exact behaviour of the
soil temperature contours, and whether steady state conditions are achieved will be
discussed in detail later in this section.
It is interesting to note that it takes a significant amount of time for the heat from the
pipe to reach the surface. This suggests that the pipe is buried too deep beneath the
surface of the soil and the possibility of moving the pipe to a more shallow depth
should be investigated to examine the effect on heat loss to the surrounds. Currently
the pipe is 1m below the surface and as shown in Figure 4.11 the radiation does not
reach 0.5m below the surface. It is proposed that the depth of the pipe should be
changed to 0.5m and then the simulation should be run again to see if there is better
heat loss from the pipe to the surface.
Shown in Figure 4.11 is the surface of the soil being heated up to a range of
temperatures between 313.6K and 324K (as it is the light green contour) at a time-step
of 13 days. This is a significant temperature and appears to be caused mainly by
radiation as the contour covers the entire surface, not just the part above the pipe. This
shows that the main mode of energy transfer opposing the transfer of heat from the
pipe to the surface is radiation from the sun. To increase the efficiency of the cycle,
radiation should be reduced to a minimum. Although this would happen when clouds
form in the atmosphere, it cannot be relied upon to provide constant protection from
the radiation and as such a means of minimisation of radiation may need to be
investigated in a future study.
The temperature contour at the end of the 30-day time period shown in Figure 4.12
illustrates contours that display the heat being conducted out by the pipe. This
represents the radial distance out from the pipe that heat being conducted to. This is
important because the effect of laying multiple pipes in different configurations has
been examined and this distance will determine the distance apart that each pipe
should be from the others. Although models of multiple pipes were not considered in
this study a distance of 4-5 diameters seem to be appropriate to minimise the
interference between the pipes.
31
Figure 4.15: Temperature history of Nodes 1 and 2 located beneath the Pipe (see Figure 4.14).
These results also showed the need to model the simulation over a period of about one
year. Specifically, Figure 4.15 shows the nodes read a constant value after
approximately 250 days. Hence, to achieve an accurate simulation, the hourly
averaged data for a given month was repeated for 12 months of simulation time. This
simulates the worst case scenario as winter months are colder than January and the
temperature below the pipe would have been slightly lower.
4.5
2D TRANSIENT LONGITUDINAL RESULTS
Solutions for the 2D longitudinal model were generated using the methods explained
in Section 3.6 and are presented below. The first result calculated was the optimum
pipe depth. The model was simulated over a 20-day period and the depth of pipe
changed. Exit temperatures were recorded for each depth that was simulated and the
average temperature loss over the 50m length of pipe were plotted against the pipe
depth. The results of this are shown in Figure 4.16.
Presented in Figure 4.17 is the temperature at the pipe exit plotted over a 20-day
period. These data will be employed to explain why the 2D longitudinal model was
simulated for only 20 days in the solution obtained for temperature versus depth
results. Details of this explanation can be found in Section 4.6.
32
Figure 4.16: Graph of Exit Temperature versus Pipe Depth
Figure 4.17: Exit Temperature vs Pipe Depth over 20 Days for a 50m long pipe
The longitudinal model was also used to obtain an estimate of the overall pipe length.
The 50m geometry in Figure 3.11 was looped multiple times to calculate a
temperature drop over 7.4km. The results of this temperature varation are presented in
Figure 4.18 for every 50m. A trend-line has been applied to this data and the equation
of the trend-line is also shown on the graph.
33
Figure 4.18: Calculated Temperature drop along a 50m length, extrapolated over a 7.4km length
of pipe
4.6
2D TRANSIENT LONGITUDINAL DISCUSSION
Through inspection of Figure 4.16 it can be seen that the temperature loss is greater
when the pipe is at a depth of approximately 100mm. At shallower depths, the
average heat loss is lower. The likely explanation for this is that the influence of solar
radiation increases with reduced depth. Since soil is a poor conductor, even a small
layer helps to insulate the pipe from this additional energy and allows the pipe to cool
better. At greater depths, the convective losses through the surface at night reduce,
owing to relatively poor conductivity of the soil. However, it is important to verify
these results with further analysis, notably one which accounts for the spacing
between pipes.
Figure 4.16 was obtained by taking the average of the pipe exit temperature on the
20th day of the cycle and plotting the derived temperature loss of the working fluid for
different depths. The 20th day was chosen as the length of time over which the
solution was found because the temperature in the pipe was found previously to reach
steady state after approximately 18 days. Figure 4.17 represents the exit temperature
fluctuations at varying depths over a time period of 20 days. This shows that the exit
temperature is not a constant value but oscillates throughout the day. As can be seen
in this figure, the fluctuations after about 18 days have an amplitude of approximately
0.1K which is deemed to be comparatively small (thus allowing steady state
behaviour in the working fluid to be assumed). Please note that the decrease in
average fluid temperature inside the pipe is the measure of the effectiveness of this
approach to provide cooling. The reduction is relatively small, ~1K at the exit of a
50m long pipe which is rather small. This is likely to be the case if we consider the
low conductivity of the soil, the lack of any fins to increase surface area and the very
hot conditions the surface of the soil is exposed to. It is a worse case scenario.
34
The fluid exit temperature was calculated for a period of 20 days as shown in Figure
4.18, and is plotted versus the pipe length. These calculations were conducted for 50m
long pipes where the exit temperature from the first run is used as inlet for the second
50m pipe. The process was repeated numerous times until the temperature was
determined for a 7400m pipe. This computationally elaborate and expensive approach
was necessary due to the length of the pipe and the length of modelling time required
to achieve steady state conditions. Unfortunately it was not possible to continue this
process until the 30C exit temperature was achieved. A curve fit however was used to
represent the data for the 7400 pipe which caused the temperature of the fluid to drop
by ~25C.
Noteworthy is that the temperature difference between the pipe and the surrounding
soil, which drives the potential for heat transfer, drops and hence the effectiveness of
a meter length of pipe drops too. Hence extrapolating the curve generated above to
longer pipes may not be accurate. Nonetheless, the curve fit is for almost half of the
temperature reduction required (70C) and a rough estimate can be generated using
this method. Thus, solving for length, and using a value of 300K (27C) for the pipe
temperature, the length of pipe required is approximately 22km. And in order to
account for the issues discussed above a conservative estimate was deemed to be
prudent and a factor of safety was introduced. Hence for cost analysis only the overall
length of pipe required for the underground cooling process was therefore set to 25km.
4.7
RESULTS OF COST ANALYSIS
On deciding the size and material of the pipe two competing factors need to be taken
into account; enhancing heat transfer from the pipe and reducing pumping cost due to
pressure drop inside the pipe. For a smooth pipe one can use an estimate of pressure
drop per meter length from a figure like the one shown in Figure 4.19. According to
the industry guidelines the water velocity in the pipe should be maintained 1.2m/s,
for pipes which are 50mm in diameter and the pressure drop 400 Pa/m. Using these
values as a guide it became quite clear that the pipe system need to be arranged in
parallel to reduce the flow speed and minimise the pressure drop. In this case and for
50mm diameter pipes the flow rate is 2.4 L/s corresponding to ~400 Pa/m.
Figure 4.19: Friction Loss for Water in Plastic Pipe (Schedule 80) (ASHRAE)
35
And since the total flow rate is 16 L/s of water the 25km length of pipe will need to be
divided into 7 parallel sections. Increasing the diameter to 75mm will increase the
flow rate per section to 5.3 L/s, decrease the pressure drop to ~300 Pa/m and the
number of parrel sections to 3. These arrangements will most certainly impact on the
exit temperature both ways. In one hand, this will lead to multiple starts and hence
high temperature differential on multiple sections and in another we may need the
length to achieve the required low exit temperature. Since the aim of this study is to
assess the feasibility of the concept, these design decision will be left for future
studies when the size and location of the system is known.
For the rest of the cost estimate section it is assumed (see Section 4.5) that the pipe
length needed to dispose of 5MW of heat is approximately 25km and the optimum
pipe depth is 0.1m. These data were used to obtain a rough estimate of costs.
4.7.1 COST ESTIMATE OF UNDERGROUND HEAT EXCHANGER

To estimate the cost of burying the pipe, a quote was obtained from Mr Geoff Pike,
Managing Director of General Trade Industries (GTI), who has considerable
experience in the area of laying pipes and pipelines. According to Mr Pike, the most
effective way of completing such an installation would be to use a grader which can
form a 100-200mm deep trench (and back-fill) 5km long in one day. Therefore a work
program consisting of two polyethylene layers and a grader driver would be able to
complete the desired 25km in five days. The most effective spacing for each pipe
would be indicated by the grader wheel span, which is nominally 3m apart. A
breakdown of the costs provided by GTI is presented in Table 4.1.
Table 4.1: Breakdown of estimated costs of burying the pipe
Activity
50mm Diameter Pipe
Machinery
Labour
Operator
Mobilisation & Demobilisation of plant
and equipment
Total Estimate
Estimated Cost
$70,000
$5,000
$7,000
$3,000
$15,000
$100,000
This indicates that the total capital cost for burying the pipe is approximately
$100,000. However, pumps are required in order to circulate the working fluid
through the pipes. Hence, in the estimate of capital costs, the cost of pumps must be
included. In order to have correctly sized pumps an estimate of the pressure head and
flow rate required of the pumps must be calculated. The volumetric flow rate may be
simply obtained by multiplying the mass flow rate (16kg/s, as stated previously) by
the average density of the fluid. The average density was deemed to be the average of
the density of the working fluid at the inlet (T=100C, =968 kg/m) and the density of
the fluid at the outlet (T=30C, =996 kg/m) of the condenser pipe. However, as the
proposal provided by General Trade Industries indicates that five pipes of 5km length
be used in parallel, which, by continuity, indicates that the mass/volumetric flow rate
36
will drop by a factor of five. With these data, the required volumetric flow rate was
deemed to be approximately 12m3/hr.
The pressure head for a 5km long DN50 polyethylene pipe is estimated at 200m water
head. This is the minimum pressure head that the pump must be able to supply in
order to circulate the water through one 5km branch of pipe. A pump may now be
sized for this application. The pump deemed to be of sufficient capacity was a
Grundfos Vertical Multi-Stage pump, Type CRI 15-17, and the cost of this pump
according to R&D Pump Equipment and Services (Pty. Ltd.), was $5,862.00. The cost
of five of these pumps is therefore $29,310.00. The relevant cost estimate is shown in
Appendix C.1.
The final capital cost estimate for the underground heat exchanger is therefore
$129,310.00.
The on-going costs of the proposed underground heat exchanger may now be
examined. The majority of the on-going costs for the underground heat exchanger will
come from the cost of running the pumps. As there are five pumps rated at 15kW each,
the total power requirement is 75kW. If the electricity cost is estimated at $0.15/kWhr
and the pumps are assumed to run at full capacity for the whole year, then the total
on-going cost for the underground heat exchanger is $98,550.00 per annum. Note that
the assumptions of electricity cost and how long the pumps are run for are not
expected to be representative of the actual situation. Nonetheless, these assumptions
are still valid as the same assumptions will be employed in the cost estimate of the aircooled heat exchanger, which makes the costs valid comparatively.
4.7.2 COST ESTIMATE OF AIR-COOLED HEAT EXCHANGER

A cost estimate for the air-cooled heat exchanger was provided by CBM Radiators
Australia. It was specified that this heat exchanger would need to dissipate 5MW of
heat energy, and cool the working fluid from 100C to as close to ambient conditions
as possible. The air-cooled heat exchanger is required to be designed for worst case
conditions, so that the plant may operate even at peak temperatures in summer. Thus,
the module is designed for ambient conditions of 50C, and a pipe exit temperature of
56C. The required flow rate for the working fluid is 28L/s. The actual design of the
radiator is described below.
The radiator design is of a horizontal flat deck type with electric fans mounted above
the cooling coils providing induced draft vertical discharge air flow. For a 50C
design ambient, CBM requires four of their largest modules mounted side by side and
plumbed in parallel to a single inlet and outlet water connection point. Each of these
modules will have 3 x 11.0 kW cooling fans. The modules will include 2.0m high
structural support stand structures for mounting to a concrete base. An approximate
foot print dimension of the entire package is 8.0m x 8.0m.
CBM radiators estimate the cost of this package to be $308,000 (including GST). To
be included in the capital cost is the cost of the pump required to circulate the water
through the exchanger. With the data provided from CBM, the pump deemed suitable
for this application was a Grundfos Vertical Multi-Stage pump, Type CRI 20-14.
37
According to R&D Pump Equipment and Services (Pty. Ltd.), the cost of this pump is
$6,343.00 (a quote may be viewed in Appendix C.1). Therefore the total capital cost
of the air-cooled heat exchanger system is $314,343.00.
The on-going costs of this heat exchanger are mostly derived from parasitic losses
associated with operating the fans and the pump. As stated above, the exchanger
requires four banks of 3 x 11kW fans, or 121kW of fan power required. The pump is
rated at 15kW, making the total power supplied to the system 136kW. Using the same
assumptions employed in the underground heat exchanger on-going cost estimate
(running full year round, cost of electricity is $0.15/kWhr), the total on-going costs
for one year of operation is approximately $178,704.00.
4.7.3 COMPARISON OF COSTS & ANALYSIS

The total and on-going costs of the underground and air-cooled heat exchangers are
presented in Table 4.2. Examination of this table indicates that both the capital and
on-going costs for the underground heat exchanger proposal are lower than those for
the air-cooled radiator. This would seem to indicate that the proposal of underground
cooling is economically feasible. This may indeed be the case, but care must be taken
to regard these costs as indicative only.
An additional substantial gain from this system, which has not been considered in this
analysis, is the increase in the thermal cycle efficiency due to the lower condenser
temperature. The reduction of the cold side temperature from 56C to 30C can yield
an increase in efficiency of ~30%. This has direct impact on the productivity of the
plant and the subsequent income potential.
Table 4.2: Cost Comparison of Air-Cooled and Underground Cooled heat exchangers
Cost Descriptor
Capital Costs
On-Going Costs (p.a.)
Underground Cooling
$129,310
$98,550
Air Cooled
$314343
$178704
Assumptions and Simplifications:

Firstly, the underground cooling system is not optimised in several respects. Further
investigation is required to qualitatively establish the most cost-effective arrangement
of pipes and pipe sizing. It was assumed for this report that heat transfer is most
important, but it may be that cost requirements are such that longer pipes, or wider
diameter pipes, are less expensive in total.
Secondly, all costs provided are rough estimates, and do not take into account the
many factors that would affect the actual implementation of both systems (such as
maintenance and repairs). These additional costs may increase the on-going costs by
10-20%. Furthermore, the cost of land was not considered in the estimation of the
capital cost of the underground heat exchanger; that is, since this system requires a
tract of uninhabited land in which to be buried, the costs associated with this may
need to be considered in future estimates. However, for the purposes of this report,
38
and with the absence of other data, it will be assumed that the cost of the land for the
heat exchanger is relatively low and will not be included.
Finally, the costs are based on theoretical predictions only, particularly for the
underground cooling case. Actual experimental testing is required to validate the
concept, and qualitatively prove that such a system would operate in the manner
indicated. Although the finite element solutions are as accurate as possible given the
inherent assumptions in the modelling, as well as software limitations, it would be
prudent to perform the experimentation mentioned above. In the case of the air-cooled
heat exchanger, it is assumed that the cooling of the working fluid during peak
conditions is adequate for the power cycle. Were this heat exchanger to be
implemented, significant efficiency losses would be expected during these conditions,
which may be unacceptable. It may be that the air-cooled system requires some
augmentation to achieve more effective cooling.
4.8
EXAMINATION OF FEASIBILITY
From the results given previously it is possible to estimate the overall feasibility of the
project. The finite element models show the technical feasibility whilst the cost
analysis predicts economic feasibility of this project. This section will present a
detailed examination of the feasibility of each of these analyses.
Before any analysis of cost may be undertaken, the technical feasibility of the
proposal of underground cooling must be established. The results from the 2D lateral
model show that the buried pipe will eventually reach an overall steady state
condition. The working fluid in the pipe will reach this condition at approximately 20
days, but the soil surrounding the pipe may take up 250 days to achieve steady state.
This indicates that an energy balance has been attained, whereby heat from the pipe is
being transferred to the surrounding atmosphere, and is not just heating up the soil
continually. This result indicates that the concept is potentially feasible. The technical
feasibility of the proposal is confirmed by results obtained from the 2D longitudinal
model. These indicate that the working fluid will likely cool to the design exit
temperature of 30C in a non-infinite length of pipe. The length of pipe deemed to be
sufficient for this process is ~25km. Therefore, it can be concluded that the concept of
underground cooling for power-cycle design is expected to be technically feasible. In
other words we are able to use the soil for thermal storage during the day and to
dispose of the heat during the night. We are also able to provide a constant lower exit
temperature of the coolant 24 hours of the day.
The economic feasibility was examined by contacting relevant industries and
obtaining quotes for the cost of laying the underground cooling system and an
equivalent air cooled heat exchanger. The cost analysis shows that both the capital
and on going costs of implementing an underground cooling system are less than the
equivalent air cooled system. For the implementation of the underground cooling
system, it is estimated that $185,033.00 will be saved in capital costs, and $80,154.00
(per annum) will be saved for the on-going costs. Thus, since both capital and ongoing costs are lower for the underground heat exchanger system, it may be stated
that the concept is expected to have a lower equipment cost. The effect on the
performance of the geothermal power station has not been calculated, and this will
39
probably have a greater impact on economic viability than the costs of the cooling
plant. Nevertheless, at times of peak ambient temperature, the above underground
system is calculated to cool to a lower temperature by 20C, which will mean higher
output at that time. At the same time, it will not provide as much cooling during
periods of lowest ambient temperature, but will provide a much more uniform
condenser temperature, and hence consistent output from the power station.
40
SUMMARY AND CONCLUSIONS
This study assessed the technical and economical feasibility of using soil as mean to
store energy during the day and dispose of it during the night in the Copper Basin
region of South Australia. Multiple approaches were used to calculate the rate of heat
transfer and the viability of the concept. Analytical as well an advanced
computational technique, namely Finite Element Analyses, (FEA) were employed in
this study. It was estimated that the optimal depth to burry the heat exchanger pipes is
0.1 m. This depth was arrived at through modelling the soil absorption of the heat
dissipated from the pipe and the impact of the sun radiation, air temperature and wind
speed on heat transfer from the soil to the atmosphere.
Further modelling of the surrounding soil temperature has estimated that the spacing
between the pipes should be at least 4-5 pipe diameters, that the pipe exit temperature
fluctuates by less than 0.1K and that a steady state is achievable within 20 days of
continuous operation.
The two dimensional longitudinal model helped us estimate the length of pipe
required to dissipate 5MW of heat by cooling a working fluid (Water) from 100C to
30C. The model revealed that an approximate length of 25km of 50mm diameter
DN50 polyethylene pipe is required to achieve the required heat transfer.
A cost analysis was performed for a simple design and yielded a rough comparison
between the underground system and an equivalent air cooled system. This compared
both installation costs and on-going costs per annum. It was found that the
underground cooling system would have lower capital and on-going costs than an
equivalent air-cooled heat exchanger system. For the implementation of the
underground cooling system, it is estimated that $185k will be saved in capital costs,
and $80k (per annum) will be saved for the on-going costs. The system is also
expected to yield a more constant power station output, with higher output than the
air-cooled system during periods of maximum ambient temperature, and conversely
during periods of minimum output, although these values were not calculated. This
analysis suggests that the underground system is likely to be economically feasible.
The overall concept of underground cooling was thus expected to be feasible.
Nevertheless, it is important to note that the model employs a number of simplifying
assumptions, and further model development and validation are required.
41
FUTURE WORK
The results conclusively show that the idea of utilising the thermal properties of soil
to act as a heat exchanger is scientifically possible. The models shown in this report
have not been optimised in a way to give the maximum heat transfer. An in depth
study could be undertaken to optimise the design such as the use of finned tubes in the
models instead of just a standard pipe. The use of finned tubes would significantly
increase the heat transfer from the pipe to the soil, but would be more expensive to
purchase and install. A cost-benefit analysis would help to establish whether the use
of finned pipes would be advantageous. In addition the suitability of the piping system
when other coolants are used will also need to be understood and accounted for.
Another method to increase the heat transfer would be to surround the pipe in a more
thermally conductive material, such as a different mix of soil with a higher
conductivity. This would have significant effects on the pipe depth and overall length,
and would need to be investigated thoroughly. The cost of using a different soil infill
would need to be examined, particularly the transport costs, as the soil would need to
be transported to the geothermal power plant location in remote northern South
Australia.
The report showed that solar radiation has a significant effect on the behaviour of the
model. Minimising solar radiation on the pipe through shading devices may be of
significant benefit, and could shorten the length of pipe required. A cost-benefit
analysis would again be important in establishing the viability of this idea. Similarly,
the heat transfer from the soil can also be enhanced through additional exposed fins
during the night to accelerate the energy transfer from the soil to the atmosphere.
A physical test needs to also be carried out to verify the findings of this project and
demonstrate practicality of the project. It is desirable to conduct the tests for a
significant time period (probably 1 month or so) during summer and at close location
to the actual sites. Alternative laboratory type experiments are possible too albeit with
additional cost. Such tests would help determine pipe depth and overall length of pipe,
which could be compared to the data in this report.
As the case study in this report only considered water as a coolant, applicable to flash
cycles and without considering the financial benefits due to improved thermal
efficiency at peak load times. Hence, it is imperative to assess the benefits for a real
cycle and for a specific location. And also to develop a financial model which takes
into account the electricity prices during the day and the potential financial benefits in
comparison with air cooled systems.
42
REFERENCES
Abu-Hamdeh, N. H., & Reeder, R. C. (2000). Soil Thermal Conductivity: Effects of

Density, Moisture, Salt Concentration, and Organic Matter, Soil Science Society of
America Journal , Vol. 64, No. 4, pp1285-1290.
ANSYS. 2007, ANSYS v11.0 Release Documentation, v.11.0, computer program,
ANSYS Inc., Pittsburgh, PA, USA
ANSYS Training Manual 2005, ANSYS Inc., Pittsburgh, PA, USA
Burdon, A 2006, 'The Strzelecki Track', Australian Geographic, Issue 82, pp42-57
Crawford, M Kays, W & Weigand B (2004), Convective Heat and Mass Transfer,
Mc-Graw Hill Professional.
Australian Government, Department of Environment, Water, Heritage and the Arts,
2001, <http://www.environment.gov.au/education/publications/greenhouse.html>,
viewed 28th April 2008
DiPippo, R. (2005). Geothermal Power Plants. Oxford, UK, Elsevier.
Energy Efficiency and Renewable Energy (EERE) 2008, Statistics for SA Oodnadatta
Airport, Department of Energy, USA
Feeley, TJIII, Green, L, Murphy, JT, Hoffmann, J, Carney, BA 2005, Department of
Energy/Office of Fossil Energys Power Plant Water Management R&D Program,
<http://www.netl.doe.gov/technologies/coalpower/ewr/pubs/IEP_Power_Plant_Water
_R&D_Final_1.pdf>, viewed 28th April 2008
Geodynamics Limited, viewed 28th April 2008,
<http://www.geodynamics.com.au/IRM/content/hfr_hfrexplain.html>
Hays, GC, Ashworth, JS, Barnsley, MJ, Broderick, AC, Emery, DR, Godley, BJ,
Henwood, A, Jones, EL 2001, The importance of sand albedo for the thermal
conditions on sea turtle nesting beaches, Oikos, Vol. 93, No. 1, pp 87-94
Kose, R 2007, 'Geothermal energy potential for power generation in Turkey: A case
study in Simav, Kutahya', Renewable and Sustainable Energy Reviews, Vol. 11, No.
3, pp505
Langman, A.S., Nathan, G.J., Ashman, P.J. and Battye, D. (2008) Preliminary
Investigation of the Mechanical Systems Needed to Extract Geothermal Energy from
a Typical South Australian Site, Report to: Electricity Supply Industry Planning
Council, Consulting Report MECHTEST MT-0949, School. Mech. Eng., University
of Adelaide.
Lockwood, John G (1974). World Climate, Edward Arnold Ltd, London
43
Marra, William A 2006, Heat Flow Through Soils and Effects on Thermal Storage
Cycle in High-Mass Structures, Journal of Aerospace Engineering, Vol. 19, No. 1,
p55-58
Mills, A. (1999). Heat Transfer 2nd Edition. New Jersey: Prentice Hall
Moran, M., & Shapiro, H. (2004). Fundamentals of Engineering Thermodynamics.
Hoboken: John Wiley and Sons
Munson, B Okiishi, T & Young D (2005), Fundamentals of Fluid Mechanics 5th
Edition, Wiley.
Nassar, Y, El Noaman, A, Abutaima, A, Yousif, S, Salem, A 2005, Evaluation of the
underground soil thermal storage properties in Libya, Renewable Energy, Vol. 31, No.
5, pp593598
Pavelka, M., Acosta, M., Marek, M., Kutsch, W., & Janous, D. (2006). Dependence
of the Q10 values on the depth of the soil temperature measuring point, Plant & Soil ,
Vol. 292, No. 1/2, pp175
Perry, K, Ristaino, B, Wu, Y 1995, 'Estimating temperature of mulched and bare soil
from meteorological data', Agriculture and Forest Meteorology, Vol. 81, No. 3,
pp299-323
Ricketts, C, Myint, M, Nirmalakhandan N & Gadhamshetty V 2006, Improving AirCooled Condenser Performance in Combined Power Cycle Plants, Journal of Energy
Engineering, Vol. 132, No. 2, pp81-88
Smith, D 2002, 'Evaporative Cooling Increases Summer Capacity at Geothermal
Plant', Power Engineering, Vol. 106, No. 2, pp84
44
APPENDIX A
HEAT CONVECTION COEFFICIENT CALCULATION
(A.1)
Re = V L /
Where,
Re = Reynolds Number
V = Averaged Wind Velocity = 5.47m/s

L =Unit Length=1m
=Kinematic Viscosity=1.511e-5m/s
These values were entered into Equation A.1 and solved to give a Reynolds Number
of 3.62E-5.
This implies that the flow is turbulent and as such the Nusselt Number, Nu, can be
determined by the following equation[Mills]:
Nu = 0.037 Re 0.8 (Pr)1 / 3
(A.2)
Pr=Prandtl Number=0.713
Solving Equation A.2 gives a Nusselt Number of 925.46.
This value can be used to calculate the heat convection coefficient, h, by the following
equation[Mills]:
Nu =
hc L
k
(A.3)
k=heat transfer coefficient=0.0257W/mK

Rearranging Equation A.3 and solving gives a heat convection coefficient of
23.6W/m2K.
45
APPENDIX B
DATA USED IN FEA TRANSIENT MODELS
Figure B.1 Average hourly temperature in C for a year at a weather station in the town of
Oodnadatta [EERE, 2008]
46
Figure B.2 Average daily Global radiation in kW hr/m2 for a year at a weather station in the
town of Oodnadatta [EERE, 2008]
47
APPENDIX C
COST ANALYSIS QUOTES
48
49
50

Heat

Загружено:

Сведения о документе

Авторское право

Доступные форматы

Поделиться этим документом

Поделиться или встроить документ

Параметры публикации

Этот документ был вам полезен?

Это неприемлемый материал?

Авторское право:

Доступные форматы

Heat

Загружено:

Авторское право:

Доступные форматы

TIG 6.

Feasibility of Underground Cooling for

A/Prof. Bassam Dally, Prof. Gus Nathan

This project is supported by the Primary Industry and Resources in

Enthalpy, convection coefficient

Depth of pipe underground, length of pipe

heat flux due to conduction

Heat flux due to radiation

Heat flux due to convection

Heat flux emitted from the soil

Ambient air temperature

Inlet water temperature

Outlet water temperature

Diameter of the pipe

Mass flow rate

Absorbed solar energy

1.2 AIMS OF PROJECT

HOT FRACTURED ROCKS

Hot Fractured Rocks (HFR) technology is a recent development in geothermal power

SOIL HEAT STORAGE

2.2.1 IDENTIFYING SOIL PROPERTIES

2.2.2 HEAT BALANCE

2.2.3 COMPARISON OF EXPERIMENTAL & MEASURED RESULTS

2.3 ESTIMATING SOIL TEMPERATURE

Perry, Ristaino and Wu (1995) have developed a detailed approach to calculate

Figure 3.1 Schematic diagram of a pipe buried in soil

For the present comparative assessment, it is sufficient to assume a typical

In investigating a complex problem it is good practice to initially reduce the problem

(a) Section A-A

(b) Section C-C

Figure 3.3 Schematic diagram of the 1-D model

STEADY STATE 1D VALIDATION MODEL

3.3.1 SIMPLIFIED MODEL

Figure 3.4 Schematic of verification model

Table 3.1: Soil and air thermal properties

Table 3.2: Temperature data and other parameters

Expanding Equation 3.1, leads to:

3.3.2 FINITE ELEMENT MODEL

Figure 3.5: 1D ANSYS line model geometry

TRANSIENT 1D LINE MODEL

3.5.1 CONSTRUCTION OF THE FINITE ELEMENT MODEL

Shown in the reconfigured geometry is a pipe of diameter d buried in soil at a depth l

Figure 3.8: Lateral geometry with applied mesh

3.6.1 CONSTRUCTION OF THE FINITE ELEMENT MODEL

Figure 3.11: Finite element model of the Longitudinal Cross Section

Figure 3.12: Close View of Nodes and Elements

To compute the convective heat transfer process, ANSYS requires a mathematical

Nu D = 0.023 Re 0D.8 Pr 0.4

Fully turbulent flow in a pipe may be characterised as turbulent if the Reynolds

To perform a cost analysis, industry experts with relevant experience in cost

4 RESULTS AND DISCUSSION

1D TRANSIENT MODEL RESULTS

DISCUSSION: 1D TRANSIENT MODEL

2D TRANSIENT LATERAL MODEL RESULTS

Figure 4.13: Contour plot corresponding to approximately 1 year

Figure 4.14: Two nodes selected for analysis

2D TRANSIENT LATERAL DISCUSSION

2D TRANSIENT LONGITUDINAL RESULTS

Figure 4.16: Graph of Exit Temperature versus Pipe Depth

2D TRANSIENT LONGITUDINAL DISCUSSION

RESULTS OF COST ANALYSIS

4.7.1 COST ESTIMATE OF UNDERGROUND HEAT EXCHANGER

Table 4.1: Breakdown of estimated costs of burying the pipe

4.7.2 COST ESTIMATE OF AIR-COOLED HEAT EXCHANGER