Numerical simulation studies of ihe geothermal resource in singapore 2

Figure 1.1: Plate tectonic map of the world showing locations 1-50 of selectedsubmarine and terrestrial high temperature hydrothermal sites after [2]Some examples of the geothermal surfa

Chapter 1

Introduction

Geothermal energy development in Singapore is more feasible than was apparent acouple of decades ago In the past, the fact that Singapore is a small island with norecent volcanic activity suggested that development for conventional geothermal energywould be uneconomical Nevertheless, the presence of natural hot springs and apparenthigh heat flow in Singapore give rise to a potential for geothermal energy development,most likely with an unconventional method Present day uncertainty in fossil fuelsupply, its price and its negative environmental impacts have made the geothermalenergy and other alternatives, like solar and nuclear energy, become more attractivefor Singapore Utilization of solar energy is still costly and it also requires a sufficientlylarge surface area, which is a challenge for a small island like Singapore Development

of nuclear energy remains uncertain, especially after the incident at Fukushima nuclearpower plant in Japan in 2011

1.1 An Overview of Geothermal Technology

Geothermal technology makes use of heat energy generated deep in the earth for powerproduction and other forms of heat-related applications It uses geothermal energywhich is a clean and renewable source of energy Radioactive decay of various iso-topes from the earth’s mantle and core represents the primary source of the exploitedgeothermal energy [1] Conventional geothermal resources are often found in locationsassociated with high volcanic or tectonic activities, such as along the tectonic platedivergent or convergent boundaries Hydrothermal activities on the surface are goodindicators for the nearby subsurface geothermal resources (see Figure 1.1 for worldwidesite of hydrothermal activity)

Figure 1.1: Plate tectonic map of the world showing locations (1-50) of selectedsubmarine and terrestrial high temperature hydrothermal sites (after [2])Some examples of the geothermal surface features are geysers, hot pools, fumaroles,hot springs, mud pools, steaming ground, and so on (Figure 1.2) Monitoring of the

mass and heat flows, and the fluid chemistry of these surface features are useful foridentifying the scale of the potential geothermal resource underground [3].

Figure 1.2: Some examples of geothermal surface features: a Mud pool near Kraflavolcano, Iceland (modified after [4]); b Steaming ground at Karapiti, New Zealand;

c Strokkur geyser, Iceland (after [5]); d The biggest fumarole in Da-You-Keng,Taiwan (modified after [6]); e Black smoker vent at East Pacific Rise 21 ◦N

hydrothermal field (after [2])Geothermal power plant is a baseload power source, i.e it can operate continuously

at up to 98 % capacity because it has a constant source of fuel (heat) and it requires avery low downtime for maintenance [7] It has far smaller greenhouse gases emissionsper unit of electricity generated compared to a conventional thermal power station,e.g coal or gas power plant [8] Some examples of geothermal power plants are shown

in Figure 1.3

At medium and high heat content (enthalpy) conventional geothermal resources(generally > 100 ◦C), highly pressurized fluid from natural geothermal systems isbrought to the surface through wells that vary in depth from a few hundred metres

to 2.5 kilometres [12] Hot fluid brought to the surface is separated into steam and

Figure 1.3: Examples of geothermal power plants: a Larderello power station, Italy(after [5]); b The 47.4 MWe Germencik power plant, Turkey (after [9]); c The 3MWe binary ORC power plant at Landau, Germany (after [10]); d The Geysers,

California, USA (after [11])brine The steam is used to turn the rotors of turbines, which in turn spin generators

to produce electricity The brine is used to produce more steam (e.g by flashing) forelectricity generation, to supply heat for other relatively low temperature applications,

to be reinjected back to the ground through reinjection wells (Figure 1.4), or to bedrained away Injected fluid is useful for maintaining reservoir pressure in the laterstages of production [13] It can also be used as the source of recharge fluid to berestored, reheated and later re-extracted again through the production wells

Low enthalpy conventional geothermal resources are ubiquitous but more suitablefor direct uses, like health spas, heating greenhouses, geothermal prawn farming (e.g

on the banks of Waikato River in New Zealand), space conditioning of residential andcommercial buildings by installing heat pump systems, cooking, industrial heating,and so on Electricity can also be generated from this type of resource by using binaryplant technology, for example: the 0.21 MWe geothermal power plant in Neustadt-Glewe, Germany, uses the Organic Rankine Cycle (ORC) to generate electricity from

Power Plant

Production Well

Magma

Injection Well

GEOTHERMAL POWER PLANT

Figure 1.4: Schematic of geothermal power plant production and injection wells

(modified after [14])

98◦C fluid with a flowrate of 35 L/s [10] The energy-depleted brine is either drainedaway or reinjected back to the ground Current practice is to reinject the brine back

to the subsurface geothermal system in order to minimize release of geothermal fluids

on to the surface of the earth

The worldwide installed capacity of geothermal plants in 2010 is shown in Figure1.5, with total installed capacity of 10,898 MWe The top five countries in terms ofinstalled capacity are: USA (3098 MWe), Philippines (1904 MWe), Indonesia (1197MWe), Mexico (958 MWe), and Italy (843 MWe) The forecast total worldwide in-stalled capacity for 2015 is 19.8 GWe, which is almost double the installed capacity in

2010 (see [15] for more statistical data) Such a high prediction is supported by theadvancing technology in binary plants, which enables geothermal power production

at sites with low to medium enthalpy geothermal resources It also improves powerproduction efficiency at high enthalpy geothermal resource sites by allowing furtherpower production for the 100-180◦C geothermal water that otherwise would have beeninjected back to the system

Figure 1.5: Worldwide geothermal power plants installed capacity in 2010 (after [15])1.2 A Brief History of Geothermal Energy

1.2.1 Conventional geothermal resource

The use of geothermal heat can be traced back to very early times when it was usedfor direct-use applications, like therapeutic bath and cooking In the wild, naturallyoccurring hot pools are used by some monkeys (Figure 1.6b) for survival against thewinter cold temperature The first human use of geothermal heat in North Americaoccurred more than 10,000 years ago with the settlement of Paleo-Indians at hot springs[16] In the Roman times, steam and hot water springs were used for hot water andbathing (Figure 1.6a)

In modern times, geothermal energy was first exploited by Prince Piero GinoriConti for a power generation demonstration by using emerging steam to drive a small

Figure 1.6: Use of geothermal heat: a Roman bath at Bath, England (after [5]) ; b.

Snow monkeys soak in hot pools in Japan (after [17])turbine to provide electricity for five incandescent light bulbs in Larderello, SouthernTuscany, Italy, in 1904 [18] Nine years later the first geothermal power plant pro-ducing 250 kWe was constructed at Larderello In 1958, the second geothermal powerplant was built in Wairakei, New Zealand In 1960, the third geothermal power plantproducing 11 MW electricity was in operation in the Geysers, USA The world totalinstalled capacity has been increasing since then (Figure 1.7)

0 5,000 10,000 15,000 20,000

YearsFigure 1.7: World geothermal installed capacity from year 1950 to 2010 and a

forecast capacity at 2015 (data after [15])

1.2.2 Overview of EGS

Conventional geothermal resources rely mainly on the availability of permeable rock,and a sufficient amount of groundwater as the medium to transport the heat energy.Permeable rock combined with groundwater and a relatively shallow heat source results

in hydrothermal a system Such systems are mainly found at the locations shown inFigure 1.1 At locations far away from the tectonic or volcanic activities, either therock is not permeable enough, the groundwater is scarce, or the heat source is toodeep for economical exploitation However, in some areas like the caprock or themargins of hydrothermal systems [19], the heat source is close enough to the surface,but the rock is not permeable enough Fracturing the rock (via hydraulic, thermal,and/or chemical stimulations) increases its permeability, and pumping of water fromthe surface through these artificial fracture networks allow for heat extraction fromthe relatively shallow and hot dry rock This technique enhances the reservoir, andhence the name Enhanced Geothermal System (EGS), also known earlier as Hot DryRock (HDR) system The general concept is shown in Figure 1.8, where cold water

is pumped via an injection well to flow through the artificially fractured rock As itflows, heat is transferred from the hot rock to heat up the flowing water The heatedwater is then extracted through the production well and used for power generation

In 1977, Los Alamos established the first operational HDR circulation loop atFenton Hill A series of experiments were carried out but the HDR system failed toproduce power at a commercial scale Nevertheless, the experiments included the firstgeneration of electricity from a HDR system, a modest 60 kW of electricity using abinary turbine-generator [21] Since then, many other experiments with HDR systemshave been carried out at other places (Figure 1.9), like Rosemanowes, Hijiori, Soultz,Cooper Basin, and so on (see [20] for more) A general lesson from these experiments

Figure 1.8: Schematic of two-well EGS in a low-permeability crystalline basement

formation (modified after [20])

Figure 1.9: Evolution of global EGS projects with estimated electrical power output

per production well (after [20])

is that in order to minimize runaway water losses and short-circuiting when creatingthe stimulated reservoir rock, the use of low-pressure stimulation (hydroshearing) ispreferred to high-pressure hydraulic fracturing [21].

A MIT study sponsored by the U.S Department of Energy in 2006 [20] concludedthat EGS in the United States could provide 100,000 MWe or more in 50 years Theprojection was based on a review of past EGS projects and realization of performancecriteria that include a thermal drawdown in the production well of no more than 10

◦C over 30 years, and a flow of 50 kg/s The U.S Department of Energy (USDE),Ormat and GeothermEx recently (April 2013) announced that they have successfullyproduced 1.7 MWe from EGS technology, which marked the first EGS project to beconnected to the US electricity grid [22]

1.3 Site Examples

1.3.1 Conventional geothermal fields: Indonesia

Indonesia is situated on the Pacific ‘Ring of Fire’ (see Figure 1.1 point 17-19) and

is estimated to have approximately 28,000 MWe of potential conventional mal resources (considered to be the largest in the world) [23] Exploration for high-temperature geothermal resources in Indonesia has been carried out since 1970 [24]

geother-By year 2000, at least 6 geothermal fields had been developed with a total capacity of

800 MWe In 2010, Indonesia had 1,197 MWe of geothermal generation in operation.During the World Geothermal Conference 2010 in Bali, President Susilo BambangYudhoyono said that Indonesia aimed to be the world’s leading geothermal nation by

2025 [25]

Some of the largest geothermal resources in Indonesia are found in Java and

Suma-tra Islands This is mainly because these two islands are located in the major geologictectonic structure that controls the location of most of the volcanic and geothermalactivities on the islands.

In Java island, most of the developed resources are located within 200 km from thecapital city, Jakarta The largest developed field is the Awibengkok geothermal field,also known as Salak, which sustains 377 MW of electricity generation Together withDarajat field (135 MWe), the combined capacity of these two projects makes Chevronthe largest geothermal operator in Indonesia [26] Awibengkok is a water-dominatedreservoir with temperatures of 235 to 310 ◦C Other developed geothermal projects inIndonesia include [27]: Lahendong (20 MWe), Dieng (60 MWe), Wayang-Windu (110MWe), Kamojang (140 MWe), and a 2-MWe project at Sibayak (at Sumatra Island)

1.3.2 EGS: Soultz, France

The Soultz geothermal project has several main wells: three deviated, 5 km deep wellsdrilled from the same platform, a 2 km deep observation well, and a 3.6 km deepexploration well that can be used as an additional reinjection well (Figure 1.10) Theexperimental project is carried out with France, German and Swiss governmental andEuropean funding [28] The project’s original aim was to design large industrial units(> 25 MWe) with multi-well EGS method However, no additional wells was drilled atSoultz On the ground surface (Figure 1.11), ORC binary power plant was built with

a net capacity of 1.5 MWe

Injection testing in the Soultz project suggested that fracturing (resulted fromhigh-pressure injection) was diffuse rather than concentrated in discrete fractures TheSoultz project also demonstrated that large fractured volume can be created repeatedly

in rock containing pre-existing natural fractures that are ready to fail in shear [20]

Figure 1.10: Schematic S-N cross section through the Soultz wells GPK2 and GPK4are production wells, GPK 3 is the reinjection well GPK1 can be used as additional

re-injection well, and EPS1 is the observation well (after [28])

Figure 1.11: The Soultz geothermal power plant: the ORC power unit in the back,the three geothermal wells in the middle, and the cooled geothermal loop in the front

(after [28])

1.3.3 EGS: Cooper Basin, South Australia

Geothermal projects in Australia is unique in a way that they are mainly driven byprivate sector, unlike projects in other countries that are mainly driven by govern-ment sector Several private companies that involve in geothermal projects in SouthAustralia are Geodynamics, Petratherm, and Green Rock Energy

Cooper Basin in South Australia has substantial oil and gas reserves Oil ration in this area encountered high-temperature gradients and intersected graniticbasement with high abundance of radiogenic elements This information led Geody-namics to drill their first well (injection well Habanero-1) to a depth of 4,421 m withbottom-hole temperature of 250◦C in 2003 [20, 21] Series of stimulations were carriedout at Habanero-1 to form fractured volume that cover a horizontal pancake-shape area

explo-of approximately 3 km2 Habanero-2 was later drilled and it encountered fractures at4,325 m depth In mid 2005, flow from Habanero-2 was tested, and a flow rate of 25kg/s was achieved with a surface temperature of 210 ◦C

In 2008, Habanero-3 was successfully drilled at 560 m NE from Habanero-1 Thetwo wells was successfully connected by a stimulated fracture system located at a depth

of 4,250 m where rock temperature is 247 ◦C [29] A six-week close-loop circulationtest enabled Geodynamics to announce its ‘Proof of Concept’ in March 2009 [30]

In Dec 2012, Geodynamics announced that it has successfully carried out a major14-day stimulation at Habanero-4 well [31, 32] In May 2013, Geodynamics announcedthat the company’s 1 MWe Habanero Pilot Plant has been successfully commissioned[33] This marked the first EGS generated power in Australia This test was success-fully completed and the plant was shut-downed on Monday 7 October 2013 [34] Theplant was last operating at 19 kg/s and 215 ◦C production well head temperature,which exceeded the expected modelled values Further data analysis and further trials

are planned to be carried out It should be noted that this project has experiencedmany difficulties and is not yet economically viable at the present However, datafrom the pilot plant trials will be used for feasibility study for future small and initialcommercial scale EGS projects in Cooper Basin.

1.3.4 EGS: Newberry, Oregon

The project is located on the western flank of the Newberry Volcano 25 miles southeast

of Bend in the central Oregon, USA The EGS demonstration in the Newberry crater iscarried out by AltaRock Energy (see [35]) The project is expected to cost an estimatedUSD44 million, while similar scale conventional geothermal projects typically costbetween USD5 and USD20 million The Newberry project is one of the most ambitiousEGS projects to date It is expected to generate ten or more megawatts of electricitywhen it is completed, perhaps in 2015

In this project, the hydroshearing method is used to create a hot water reservoir indeep rock more than 2.9 km underground The hydroshearing method is claimed to bedifferent from hydrofracturing method used routinely in the oil and gas industry [36].The hydrofracturing method uses highly pressured water (up to 70 MPa at surface)

to fracture the rock at depth to increase the rock permeability Hydroshearing, on theother hand, requires a much lower pressure (about 16.5 MPa) to initiate hydroshearing

at the natural fractures in the rock When water is injected with sufficiently high sure, the existing fractures are opened and the rock will slip slightly The irregularitypresent in the rock will maintain the opening while the rock slips The aperture of theopening is about 1 to 3 mm to allow for water or steam to flow to the well bore.Increasing the fluid pressure downhole will decrease the normal stress across thepre-existing faults and fractures The direction and magnitude of the least principal

pres-stress can be determined by a mini-frac test and borehole televiewer run just belowthe casing level [37].

In Jan 2013, AltaRock announced that it had successfully created multiple lated zones from a single wellbore [38, 39] These multiple stimulated zones dramat-ically increase the flow and energy output from a well The overall effect will be tolower the cost of geothermal energy production by as much as 50 % If AltaRock’sexperiment succeeds, that could accelerate geothermal development globally At themoment, there is no commercial-level EGS plant that exists yet

stimu-1.4 The Singapore Context

The high heat flow and the existing hot spring suggest that Singapore has a potentialgeothermal resource In 2009, Oliver [40] proposed a concept utilising the EGS method

to tap heat from the potential subsurface hot rock at depth (e.g 150 ◦C at 2 kmdepth) Possible applications of the geothermal energy in Singapore include: electricitygeneration, which is likely to require a water temperature of 150 ◦C or more; districtcooling using an adsorption chiller, which requires 90 ◦C water [41, 42]; industrialprocess heating such as the heating in water desalination plant, requiring 60 - 85 ◦Cwater [43]; and/or even their combination to form a cascading application system.With the proposed EGS concept and the possible applications, a computer modelfor the geothermal reservoir will be developed in order to assist in pre-feasibility study,

in estimation of the resource size, depth, temperature, and in prediction for ment scenarios, e.g placement of wells, well capacity, and service life

develop-1.5 Objectives

The objective of the study is to set up a numerical model for the geothermal reservoir

in Singapore The model is calibrated to match the observed and expected naturalstate conditions The simulation results produced by this model can be used forassisting work on Singapore geothermal resource development, be it to determine thewell drilling depth, location of test site, estimation of heat reserve and/or other relevantplanning aspects

The model is constructed to resemble Singapore as closely as possible Thus,input data for the model are gathered from literatures and measurements from localrock samples The thermal conductivities of Singapore sedimentary and granite rocksare measured with modified GHP method Collection of most of the sedimentary rocksamples involved collaboration with the JTC Corporation The heat flow are estimatedfrom existing heat flow map surrounding Singapore and also compared against the datafrom Chevron’s database of existing wells that are close to Singapore

The specific objective of this study is to form a model calibrated to match theexpected groundwater profile, the Sembawang hot spring location, flowrate, temper-ature and salinity The calibrated model is then used for simulations of productionwith EGS method The production simulations are carried out with both single anddual porosity models Parametric studies of the rock porosities are also carried out.Being the first study of its kind at NUS and in Singapore, the study includes severalcollaborations with the geothermal experts in the University of Auckland, and severallocal government agencies, i.e MTI, JTC and BCA

1.6 Organization of Thesis

Chapter 1 gives an overview of geothermal technology development and the objectives

of the study It begins with the motivation for the study, followed by an overview ofgeothermal technology, and then followed by a brief history of geothermal developmentaround the world Both general development for conventional method and EGS tech-nology are described Site examples for both methods are also described Literaturereviews have been divided into several parts and they are included at relevant chapters.Chapter 2 describes the thermal conductivity measurement of several Singaporerock samples, of which the results will be used in the numerical model calibration.The methodology of the measurement is described The standard Guarded Hot Plate(GHP) method is introduced, and then followed by description for the modified GHPmethod The rock samples sources and preparations, and the experimental setup andprocedures are described, and then followed by the measurement results

Chapter 3 describes the general process for geothermal reservoir modelling dure It then presents the existing data that are useful not only as the input parameterfor the model, but also useful for conceptual model development The methodology forthe numerical simulation and the simulator are then described Grid structure, bound-ary and initial conditions for the model are then described before model calibration isperformed in the next chapter

proce-Chapter 4 presents the natural state calibration processes and results Baselinemodel is first calibrated to match the initial criteria setup in the conceptual model Theprocess to improve the soil saturation, salinity and temperature profiles are discussed.The baseline model is then improved to match the hot spring salinity and flowrate,and is followed by a sensitivity study for the rock thermal conductivity measured in

Chapter 2 The study is then continued to look into the effects of geological variations

at the Jurong Formation Lastly, the chapter is concluded with a natural state modelfor Singapore geothermal reservoir

Chapter 5 describes the fracture modelling for production with EGS method lations It begins with description for grid refinement to allow for fracture modelling.Simulation results for the refined grid model are presented and checked against thenatural state calibration criteria The model is then used to simulate production withEGS method Results for both simulations with single and dual porosity methodsare then discussed The models are then used to study the effects of production ratevariations and rock porosity variations Finally, the production simulations study isconcluded

simu-In Chapter 7, the study is concluded and some findings from the studies are putforward Finally, in Chapter 7, discussion and recommendations for the model aredescribed

Appendix A presents the core logging of two boreholes in Jurong Formation region.Appendix B presents publications that are authored/co-authored by the author

In SI unit, it is measured in W/mK, which is W/m2 divided by K/m Measured values(W/mK) of some common materials are [44]: air - 0.024; styrofoam - 0.033; timber -0.14; gasoline - 0.15; water - 0.58; cement mortar - 1.73; aluminium - 250.

Rock thermal conductivity is one of the required parameters to model the mal reservoir It affects the rock temperature gradient Given a fixed heat flow, a rockwith a low thermal conductivity would have a high temperature gradient, and viceversa

geother-Examples of rock thermal conductivity values can be seen in the measured rockproperties of North-Eastern Tasmania [45], which has relatively similar petrologicalproperties to Singapore rocks The thermal conductivity increases from finer grained

laminated shales, 1.9 - 3.2 W/mK, to siltstones and mudstones, 2.8 - 4.2 W/mK,and then to high quartz content sandstones, 4.2-5.4 W/mK Its granites have thermalconductivity of 3.0 - 3.8 W/mK.

Brigaud and Vassue [46] described in their article that thermal conductivity ofcommon sedimentary rocks can range from 1.5 to 4.5 W/mK This variation depends

on the rock mineralogy, porosity and its fluid content To a lesser extent, the value alsodepends on the rock structure: stratification, distribution, orientation, size and shape

of the components A more general trend is that non-argillaceous rocks have higherthermal conductivities than those of argillaceous ones, and that thermal conductivitydecreases with increasing porosity

where ΔT (◦C) is the temperature difference over a sample thickness, Δz (m), in zdirection It follows that power, P (Watt), is

P =−λAΔT

where A is the contact surface area (m2)

2.2.1 Guarded Hot Plate: Standard method

The Guarded Hot Plate (GHP) technique is a one of the primary steady-state methodsfor measuring thermal conductivity The method conforms to ASTM C177 [47] Theinstrument used in this study is the Model GHP-300 by Holometrix (see [48] for themanual)

GHP-300 is mainly used to determine the thermal performance of insulation andother materials of relatively low thermal conductance In the standard operating pro-cedure, the instrument requires two identical 30.5 cm × 30.5 cm rectangular samples,with thickness up to 7.5 cm Estimated test accuracy ranges from +2 % to +4 %.During a normal test it can take from 5 to 10 hours before thermal equilibrium isachieved

The basic principle of thermal conductivity measurement with GHP is a ment of temperature gradient (in z-direction) at steady state with a known powerinput to the system The sample is in contact with a main hot plate at one flat endand with a cold plate at the opposite end To guard against lateral heat flow and tomake sure that the heat flows only in the z-direction, the same temperature gradient

measure-is applied to the sample at the region outside the main hot plate region The hot plateoutside the main hot plate region is called the guard hot plate, or guard heater

A simplified schematic diagram of the GHP assembly is shown in Figure 2.1 justable electrical power input is connected to the main heater plate as the knownpower input into the system (Ptot in equation 2.5) The more the input power, thehigher the main heater plate temperature The guard heater is heated up by the con-troller to reach the same temperature as the main heater plate The guard heater isintroduced to minimize lateral heat flow within the sample at the main heater plateregion The heat flows one-dimensionally from the hot main heater plate through thesamples to the cold plates The temperature of both upper and lower cold plates can

Ad-be adjusted independently to reach the desired temperature difference The coolant isused to maintain the cold plate temperature as it receives heat from the sample

Upper Heat Sink

Upper Cold PlateUpper Sample Heat flow

Main HeaterGuard

Heater

Insulation

Lower Heat Sink

Lower Cold PlateLower Sample

InsulationHeat flow

Figure 2.1: Simplified schematic for guarded hot plate assembly (modified after [48])

Thermal conductivities of both the upper and the lower samples are averaged toget a representative thermal conductivity of the material In the manual book [48],the thermal conductivity of the sample can be calculated with the following equation

which is similar to equation 2.3,

EI = λaveSh

where EI is the main heater input power (W), λaveis the average thermal conductivity

of upper and lower samples (Wm−1K−1), S is the sample to main heater contact area(0.0232 m2), ΔT is the temperature difference across the sample (◦C) and d is thesample thickness (m)

2.2.2 Rock core samples

Jurong rock samples are retrieved from available rock cores from construction projects

at Jurong Sedimentary region, Singapore The owner is JTC Corporation, whoseparent agency is the Ministry of Trade and Industry (MTI)

Two boreholes are available for this study, i.e BH9 (at Jurong Pier Rd, depth 150m) and BH16 (at DSO Carpark A, depth 155 m) Their locations are marked in Figure2.2 Cores from both boreholes have been logged by inspection of the cores at the

Bt Timah granite Alluvium Jurong sedimentary

Sembawang hot spring

Bore hole

Figure 2.2: Locations of BH9 and BH16 at the Jurong Sedimentary region

storage site (see Appendix A.1 & A.2 for the logs) The rocks vary greatly with depth,especially for BH9 At near surface (< 40 m), the rock are loose, heavily fractured andporous Alternating layers of sandstone, siltstone and mudstone are found at greaterdepth Calcareous sandstone and calcite veins are seen at depth > 130 m at BH9.Massive rocks at BH16, 31 m to 98.5 m, are greenish in color, perhaps due to chloriticclays formed by the alteration of basic tuff At depth 119.25 to 137 m BH16, unknownblack soft material is seen, with possibly high organic content.

Other rock samples include granite and other sedimentary rocks These samplesare surface rocks from various locations (Figure 2.3) For more description of thesamples, see Table 2.1

Singapore Strait

P Sentosa

P Ubin

Bt Timah Hill (164m)

Bt Timah granite Alluvium Jurong sedimentary

Sembawang hot spring

PU

Figure 2.3: Locations of granite and other sedimentary rock samples

2.2.3 Guarded Hot Plate: Modified method

For application to this study, the standard method has to be modified to fit the standard rock sample size and also to allow for only one sample to be measured at atime (due to limited availability of the rock core sample) ASTM C1044 [49] is used

non-as the guideline for the single-sided mode menon-asurement with GHP The technique is

to place the sample at one side of the hot plate, e.g upper side, and an insulator atthe other side of the hot plate, i.e the lower side Upper cold plate can then be set

at a desired cold temperature, while lower cold plate is set at a temperature as close

as possible to the main hot plate temperature This would result in almost all theinput heat to flow through the upper sample (driven by high temperature gradient)and almost no heat to flow through the lower sample (since the temperature gradient

is negligible)

Modification for the equation to calculate the sample thermal conductivity is scribed as follows

de-Electrical power input is used produce heat power (heat per unit time) to heat

up the main hot plate The heat from the main hot plate is conducted through thesample and the insulators to the cold plates (see Figure 2.4) The total heat power

Upper Heat Sink

T2 >T1

T3 T1

T = TemperatureUpper insulator + sample

Figure 2.4: Modified schematic for guarded hot plate assembly

(Ptot) is therefore the total of heat power that passes through the main sample (Ps),heat power that passes through the upper insulator (Pui) and heat power that passes

through the lower insulator (Pli), i.e.

in the assembly box are circulated with coolant from the cooler to allow for heat to

be transferred out of the system Temperatures at various points of the sample are

CoolerGHP Assembly

Figure 2.5: Diagram (a) and image (b) of overall setup for thermal conductivity

measurement with GHP method

measured with thermocouples pasted on relevant plates in the assembly box Thesethermocouples are connected to a data logger, which translates electrical signals intoreadable numbers, and then connected to a Personal Computer (PC) where the dataare compiled, displayed and stored.

Original core Sample cut with

diamond coated blade cutter

Sample with leveled surface

Fracture during cutting due to vibration

50 mm

12 mm

Figure 2.6: Cutting of sample into sizeSandstone contains quartz which is harder than steel, so it can only be cut withdiamond coated blade cutter At the same time, the rock sample is not fully cementedthat the matrix is weak and prone to fracture if subject to too much vibration from thecutting machine Wet cutting cannot be used because the sample is easily disintegrated

in water Cutting of the sample is therefore performed with a 14 inch diameter diamondcoated cutter, with dry cutting The refining cut (smoothing and polishing of flat

surfaces) cannot be performed in the lab due to tool limitation in handling the thinsample (12 mm thick), and so the samples are sent to a granite specialist companyfor refinement Further finishing touch is carried out with silicon carbide fine-grainsand paper It is important to ensure both the flat surfaces are reasonably smoothand parallel This is to ensure good contact for heat conduction into and out of thesample.

Samples other than those obtained from BH9 and BH16 are given in Table 2.1

2.4.2 GHP measurement

Prior to measurement, thermal paste is applied on both flat surfaces of the sample

to ensure good contact (Figure 2.7) The high thermally conductive paste used iseither OMEGATHERMR

2.3 W/mK), or RS thermally conductive silicon grease (5 W/mK) Both types givesimilar result Their usage depends on their availability in the lab

Figure 2.7: Thermal paste partly applied to sample flat surface

The thermal conductivity is measured with the GHP instrument For this study,the temperature of hot plate is selected to be 40◦C and cold plate to be 20◦C Theaverage temperature is 30◦C, which is approximately the in situ temperature of thesample To select the temperature difference, three measurements (temperature dif-

Table 2.1: Granite and other sedimentary rock samples from surface

from Little Guilin

ference of 5, 10 and 20◦C) are performed, and the result shows that the thermalconductivity is near stable at temperature difference of 20◦C (see Figure 2.8) Atdifferent temperature gradients, the measured thermal conductivities differ because ofundetected error, such as heat loss through the sides of the sample when heat flowsout of the sample laterally This error is insignificant for this study Nevertheless,temperature difference of 20◦C is selected for this study since the measured thermalconductivity is near stable at this temperature difference Measurement with largertemperature difference was not taken.

the desired temperature of the hot plate (40◦C) is obtained Ideally, the temperaturedifferences should be 20◦C through the sample (above the hot plate), and 0◦C throughthe styrofoam (below the hot plate) However, such exact temperature differences arevery difficult to obtain This means the heat is not solely conducted through thesample, but some are also conducted through the styrofoam Therefore, temperaturegradient through the styrofoam is accounted for in calculation (see Equation 2.7).Measurement is carried out until steady state is reached, i.e when there are negligibletemperature changes—thus, negligible apparent λ changes-in a 0.5 hour or so period.

2.5 Experiment Results

2.5.1 Instrument validation

The first measurement is performed with GHP standard method (section 2.2.1) forstandard reference material (fibrous glass board) to check the instrument performance.The result is given in Figure 2.9, with λ about 0.037 W/mK when the measurement

is stopped Since the measured λ is already close to the expected value (0.033 W/mK

at 30◦C), the measurement is stopped to save time

2.5.2 Insulator thermal conductivity

Thermal conductivity of the insulator (styrofoam) is measured with GHP standardmethod The thickness of the styrofoam is 12 mm (similar to the rock sample thick-ness), and measurement is conducted with temperature 40◦C for hot plate and 20◦Cfor cold plate (thus expected average temperature of the styrofoam is 30◦C) Theexpected thermal conductivity at 25◦C is 0.033 W/mK [44] Though thermal con-ductivity changes with temperature, its value for 30◦C is not so much different from

30◦Cthat at 25◦C Measurement is therefore conducted with average temperature of 30◦C

to maintain consistency with measurements of other samples Figure 2.10 shows theresult, with average λ = 0.031 W/mK for the last 0.5 hour This value will be usedfor all calculations to obtain thermal conductivity of the rock samples Variations atearly times in the graph are due to adjustments of power input to reach the desiredmeasurement temperature

2.5.3 Validation for the GHP modified method

To validate the GHP modified method (section 2.2.3), a mortar is cast to form acylindrical shape similar to the rock sample (diameter 50 mm and thickness 12 mm).The expected thermal conductivity of the mortar at 25◦C is 1.73 W/mK [44] Figure2.11 shows the result, with average λ of 1.64 W/mK for the last 0.5 hour The differencecould be attributed to the slightly different mix ratio between the measured mortarand the mortar used for the reference λ Since this result is close to the expected value,the GHP modified method can therefore be used to estimate the λ value of the rockcore samples

2.5.4 Results for Jurong rock samples

Measurement graphs for all rock samples are similar to those shown in Figure 2.9

- 2.11, which have trends of large initial apparent λ variations (due to input poweradjustments) and negligible apparent λ variations towards the end of measurement(steady state) All rock samples are unsaturated and all measurements are conductedwith temperatures of 40◦C for the main hot plate and lower cold plate, and 20◦C forupper cold plate The average temperature for all samples is thus 30◦C

The results for Jurong rock samples can be seen in Figure 2.12 Results for BH16

sandstone 1 (2.2) sandstone 2 (2.1)

to kaolinite For the numerical model, the λ value of this sample can be used forthe near surface rock For calculation of λ for deep Jurong rock, this low λ value is

Figure 2.13: Highly weathered rock sample from BH9 at 28 m depth; it has traces of

light grey thermal grease that is used during λ measurement

excluded (since it does not reflect the λ of less weathered rock at depth)

A representative λ for the deep Jurong rock can be estimated by idealising the rocklayers as a system of n number of horizontal rock layers that are stacked vertically toform a sandwich of those horizontal layers with a total vertical thickness of Δztot.Each layer has different thickness and thermal conductivity If a vertical heat flow

qz is supplied at the bottom of that system, and the heat is only allowed to conductvertically through those horizontal layers, then the heat flow at every layer is the same

as the input heat flow, i.e

bore-If these two boreholes are used to represent Jurong sedimentary rock in the merical simulation, then the thermal conductivity of 2.1 - 2.9 W/mK can be used forthe Jurong rock at depth, and λ of 1.4 W/mK for the Jurong rock at near surface (orresidual soil).

nu-2.5.5 Results for other rock samples

Thermal conductivity measurement results for granite and other sedimentary rocksamples in this study are tabulated in Table 2.3

Thermal conductivities of the granite samples in this study (PU, CV and SQ) arewithin the common value range for granite (1.7 - 4.0 W/mK [44]) The QueenstownFacies sample, particularly RM, is at the low end value for sedimentary rock (λ of lowporosity sedimentary rock can be as low as 0.5 to 1.5 W/mK [50]) This is probablydue to the abundance of low conductivity clay minerals in the mudstone sample

Table 2.2: Stratigraphic layer thickness (Δz) and measured thermal conductivity (λ)

from each sample at BH9 and BH16

Table 2.3: Measured thermal conductivity of collected samples: granites and other

sedimentary rock samples

Figure 2.14: Thermal conductivity of Singapore rocksMeasured thermal conductivity of the Jurong sedimentary is 1.4 W/mK for shallowdepth, and 2.0 - 3.6 W/mK for deeper depth The representative λ values of JurongSedimentary for numerical simulation is 1.4 W/mK (residual soil), and 2.1 to 2.9W/mK (deep) Thermal conductivities of the surface rock samples are: granite 1.9

- 3.5 W/mK; Gombak gabbro 2.3 W/mK; Sentosa mudstones 0.8 - 1.6 W/mK; and

Murai slate 1.3 W/mK.