Thermal conductivity of soils and rocks from the ...

Engineering Geology 164 (2013) 131–138

Contents lists available at ScienceDirect

Engineering Geology journal homepage: www.elsevier.com/locate/enggeo

Thermal conductivity of soils and rocks from the Melbourne (Australia) region D. Barry-Macaulay a,1, A. Bouazza b,⁎, R.M. Singh b,2, B. Wang a,1, P.G. Ranjith b,3 a b

Golder Associates Pty. Ltd., Building 7, Botanicca Corporate Park, 570-588 Swan Street, Richmond, VIC 3121, Australia Monash University, Department of Civil Engineering, Bldg. 60, Clayton, Melbourne, VIC 3800, Australia

a r t i c l e

i n f o

Article history: Received 23 September 2012 Received in revised form 20 June 2013 Accepted 29 June 2013 Available online 13 July 2013 Keywords: Thermal conductivity Soils Rocks Laboratory Measurement

a b s t r a c t The thermal conductivity of soils and rocks is an important property for the design of thermally active ground structures such as geothermal energy foundations and borehole heat exchange systems. This paper presents the results of a laboratory study of the thermal conductivity of soils and rocks from around Melbourne, Australia. The thermal conductivity of six soils and three rock types was experimentally measured using both a thermal needle probe and a divided bar apparatus. Soil samples were tested at a wide range of moisture contents and densities. The results demonstrated that the thermal conductivity varied with soil moisture content, density, mineralogical composition and particle size. Coarse grained soils were observed to have a larger thermal conductivity than fine grained soils. In addition, the thermal conductivity of soils increased with an increase in dry density and moisture content. Siltstone, sandstone and basalt rock samples were tested dry and water saturated. They demonstrated an increase in thermal conductivity with an increase in density when dry. However, when water saturated, siltstone and sandstone showed no significant correlation between density and thermal conductivity; whereas a linear increase in thermal conductivity with density was observed for the saturated basalt samples. These differences were attributed to both variations in mineralogy and anisotropy of each sample. The thermal conductivity data obtained from this study provides an initial database for soils and rocks from the Melbourne (Australia) region which can serve for the design of thermo-active structures installed locally and in locations with similar ground conditions. © 2013 Elsevier B.V. All rights reserved.

1. Introduction Thermally active ground structures, such as geothermal energy foundations and borehole heat exchange systems are gaining interest in Australia due to the great potential use as an aid in tackling climate challenges and meeting legislation requirements for greenhouse gas emissions (DeMoel et al., 2010; Bouazza et al., 2011; Johnston et al., 2011; Wang et al., 2012). Their efficiency and performance are dependent on the heat transfer and storage capacity of soils and rocks in which they are embedded in. In this respect, knowledge of the thermal conductivity of local soils and rocks is essential for their design. However, information on thermal conductivity of Australian soils is scarce and feasibility design values often rely on data sourced from overseas. Measurement of soil and rock thermal conductivity can be undertaken by either laboratory or field methods (Mickley, 1951; Van Rooyen and Winterkorn, 1957; Nakshabandi and Kohnke, 1965; Penner et al., 1975; ⁎ Corresponding author. Tel.: +61 3 9905 4956; fax: +61 3 9905 4944. E-mail addresses: [email protected] (D. Barry-Macaulay), [email protected] (A. Bouazza), [email protected] (R.M. Singh), [email protected] (B. Wang), [email protected] (P.G. Ranjith). 1 Tel.: +61 3 8862 3500; fax: +61 3 8862 3501. 2 Tel.: +61 3 9905 4981; fax: +61 3 9905 4944. 3 Tel.: +61 3 9905 1956; fax: +61 3 9905 4944. 0013-7952/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.enggeo.2013.06.014

Farouki, 1986; Ewen and Thomas, 1987; Brandon and Mitchell, 1989; Abu-Hamdeh and Reeder, 2000; Ochsner et al., 2001; Dali Naidu and Singh, 2004; Chen, 2007; Abuel-Naga et al., 2008, 2009; Singh and Bouazza, 2013). Field tests tend to give a gross value of thermal conductivity, while the laboratory tests provide a point value. Laboratory methods are typically used as they are relatively inexpensive, quick and allow for greater control over the boundary conditions compared to field methods. Furthermore, laboratory tests are useful for the calculation of the length of the heat exchangers thus allowing the cost evaluation of a thermo active ground structure project to be made especially during planning stages. In some other cases such as in the case of smaller residential projects where in-situ tests are seldom carried out due to financial constraints, only laboratory tests can be used to calculate the length of heat exchangers. Laboratory approaches to measuring soil and rock thermal conductivity can be divided into two main groups: steady state and transient state. Both methods have been used extensively to study the thermal conductivity of soils. Mitchell and Kao (1978) evaluated several methods of testing soil thermal conductivity and found that transient state methods, in particular the thermal needle probe, were most suitable because of their relative simplicity and short measurement time. Jackson and Taylor (1986) found that the main advantages of transient state methods were: (1) moisture migration in response to temperature

132

D. Barry-Macaulay et al. / Engineering Geology 164 (2013) 131–138

gradients was minimised and (2) a long wait for thermal gradients to equilibrate was not required. However, transient methods such as thermal needle probes can be difficult to apply to rocks. Samples large enough to eliminate boundary effects of the needle probe are required, and contact resistance errors are created when a hole is drilled into the rock sample. Therefore, in this study a transient state thermal needle probe was used to measure the thermal properties of soil samples and a steady state apparatus was used to measure the thermal conductivity of rock samples. This paper presents the results of an experimental study on the effects that moisture content, density and mineralogy have on the thermal conductivity of six soils and three rock types from the Melbourne region. Thus, providing the necessary information needed for the design of thermo-active structures installed locally and in locations with similar ground conditions. 2. Thermal conductivity measurement methods 2.1. Thermal needle probe The thermal needle probe used in this study was a commercially manufactured probe referred to as a KD2 Pro thermal properties analyser manufactured by Decagon Devices. It is based on the infinite line heat source theory and calculates the thermal conductivity by monitoring the dissipation of heat from the needle probe. Its use in the present investigation followed the procedure described in the KD2 Pro user manual (Decagon Devices, 2006). The needle probe was heated for a time, tb (approximately 30 s) where the temperature was monitored in the needle during heating, and for an additional time of tb after heating. The final two thirds of the heating and cooling data are used in a simultaneous least squares computation which determines the thermal conductivity while removing the effects of temperature drift during computation. In this study the KS-1 probe (60 mm in length and 1.27 mm in diameter) was used to measure the thermal conductivity of the soils. The probe was calibrated prior to testing using glycerol which was supplied by the manufacturer. The manufacturer claims that the needle can measure the thermal conductivity to an accuracy of ±5% between 0.2 and 2 W/mK. The KD2 Pro calculates the accuracy of each measurement by comparing the experimental temperature data to the modelled temperature predicted by the analytical solution of infinite line source theory by Carslaw and Jaeger (1959). The difference between experimental and modelled temperature is displayed as the coefficient of correlation. Measurements with correlations of less than 0.9995 were discarded and retested. A small number of measurements on samples of moist, dense sands were outside the manufacturer's recommended measurement range. Thermal conductivities of up to 3 W/mK were measured in these samples. We consider the thermal needle probe used in this study capable of measurements up to 3 W/mK without any significant errors in these types of soils. This was backed up by the coefficient of correlation readings of above 0.9998 from the KD2 Pro in all samples tested above 2 W/mK. The accuracy of the probe was found to be influenced by contact resistance errors which were created during insertion of the needle into the soil specimen. Contact resistance errors were found to be most common in low and high density soils. In soils with low densities, insertion of the needle caused disturbance of soil which resulted in regions of poor contact between the soil and the probe. For higher densities it was not possible to push the thermal needle into the samples; in these cases a 1.3 mm diameter hole was pre-drilled in the soil to facilitate needle insertion. However, the drilling caused extra disturbance within the soil and thus regions of poor soil probe contact developed. To reduce the poor contact the needle was coated with high thermal conductivity grease (thermal grease) prior to insertion. In all cases the use of thermal grease improved the accuracy of the thermal needle probe. However, in some instances in loose soils this did not improve the accuracy to an acceptable level (coefficient of correlation of 0.9995). In these cases the thermal

needle probe was removed from the sample, reinserted in a different location and the measurement was repeated. 2.2. Divided bar apparatus A steady state method in the form of a divided bar apparatus was adopted for testing the rock samples. This was used instead of the thermal needle probe as it was impractical to insert a needle into the rock samples. The divided bar apparatus used in this study is illustrated in Fig. 1 and was designed based on devices described by Sass et al. (1984), Beardsmore and Cull (2001) and Jones (2003). The divided bar consists of two temperature controlled plates at the top and bottom of the cell. The bottom plate contains an electric heater which generates a heat source of constant temperature, while cool water is circulated through the top plate from a temperate controlled water bath. Heat flux sensors 50 mm in diameter positioned either side of the rock sample measured the heat flux flowing through the rock and the temperature gradient across the specimen. When the sample reached equilibrium the thermal conductivity was determined using Fourier's law of heat conduction as follows: λ¼

Q ΔT=L

ð1Þ

where λ (W/mK) is the thermal conductivity, Q (W/m2) is the heat flux, ΔT (K) is the imposed temperature gradient, and L (m) is the height of the rock specimen. The heat flux sensors used were manufactured by placing a 1 mm polycarbonate disc between two 3 mm aluminium discs. Holes were drilled in the aluminium discs and thermocouples inserted to measure the temperature of the disc (Figure 1). The heat flux was calculated by rearranging Eq. (1) where the thermal conductivity of the polycarbonate disc was 0.20 W/mK. In practice it is not possible to simulate pure heat flow through the sample due to radial heat losses. In the present case, the samples were insulated with polyethylene foam to minimise any radial heat losses. In addition, contact resistance errors between the sample and heat flux sensors were minimised by coating the sample surface with thermal grease and by applying an axial load on the sample to ensure that good contact was established. Heat losses were monitored by taking heat flux measurements at the top and bottom of the sample. Any difference in heat flux measurements effectively represents heat loss from the sample. The heat flux measurements recorded showed minimal heat loss occurring across the sample.

Pressure

Cool Water T1 T2

Insulation Sample T3 T4

Heater Fig. 1. Cross-section of divided bar apparatus for measuring thermal conductivity. T1–T4 represent temperature measurements from the heat flux sensors.


133

Table 1 Physical properties of soils. Sample

Basaltic Clay Coode Island Silt Fishermens Bend Silt Residual Siltstone Brighton Group clayey sand Brighton Group sand

Approximate sample depth (m)

LL

1 5 25 1 2.5 14

70 82 42 47 – –

PI

50 52 24 25 – –

The samples were first tested fully saturated and then oven dried and tested again dry. The tests were run with an average sample temperature of close to 28 °C, with a thermal gradient ΔT across the sample of between 6 and 15 °C. In order to minimise moisture migration run times were kept as short as possible with each test taking between 20 and 45 min to reach equilibrium. 3. Thermal conductivity of soils The thermal conductivity of six soils was tested at a range of moisture contents and densities. The six soils comprised: Coode Island Silt (CIS) (silty clay), Fishermens Bend Silt (FBS) (sandy clay), Residual Siltstone (RS) (silty clay), Basaltic Clay (BC) (silty clay), Brighton Group Sand (BGS) and Brighton Group Clayey Sand (BGCS). Four soils CIS, FBS, RS and BC are fine grain soils while BGS and BGCS are coarse grain soils. The soils were sampled from drill spoil during bored or CFA pile construction or from test pit excavations. The particle size distribution, Atterberg limits and specific gravity of each sample are presented in Table 1 and a summary of the mineralogical composition of each soil is presented in Table 2. Samples at different dry densities and moisture contents were prepared by static compaction for thermal conductivity testing. The soils were first oven dried at 105 °C. The fine grain soils were then ground to a powder using a mechanical grinder and coarse grain were broken up by a hammer where necessary. The soils were moisture conditioned by spraying and mixing water thoroughly to the desired moisture content. They were then sealed in air tight bags and left to cure for up to one week in order to achieve a homogenous moisture distribution throughout the soil samples. A cylinder 200 mm in length and 100 mm in diameter and a loading ram 200 mm in length and 98 mm in diameter were used to compact the soil samples. They were statically compacted in 5 even layers under a loading frame. Each soil sample was compacted at different forces of 1 kN, 5 kN, 10 kN, and 20 kN per layer to achieve a range of densities. The height of each layer was measured after compaction. The thermal needle probe was then inserted into the compacted soil sample and the thermal conductivity measured. 3.1. Influence of density The effect of density on the thermal properties of soils at a constant moisture content is illustrated in Figs. 2 and 3. The thermal conductivity of the soils was observed to increase with increasing density at all

Specific gravity

% clay

% silt

% sand

(b.002 mm)

(0.06-0.002 mm)

(N.06 mm)

2.67 2.64 2.70 2.68 2.59 2.61

60 25 31 37 20 0

40 61 48 55 19 3

0 14 21 8 61 97

moisture contents. This was expected as samples with higher densities will have more soil particles and less air molecules per unit volume which will increase heat transfer. Moreover, improved packing of the soil particles improves the inter-particle contact points and therefore will lead to better heat flow between soil solid particles. Plots of the thermal conductivity against the volumetric fractions of air and soil solids are presented in Fig. 4. The relationship was observed to be stronger with the volume fraction of air, than with soil solid particles. This was particularly noticeable for the four fine grain soils which showed a very strong linear correlation between the volume of air and thermal conductivity (Figure 3a), resulting in a coefficient of correlation R2 = 0.94. Ochsner et al. (2001) also found that the volume fraction of air showed a much stronger correlation to the thermal conductivity compared to the volume fraction of solids or volume fraction of water. They obtained a correlation of R2 = 0.93 between the thermal conductivity and volume fraction of air for four medium textured soils. The coarse grain soils in this study also showed a linear increase in thermal conductivity with decreasing volume fraction of air. However, the correlation was not as strong and the dry soils did not follow the same trend as the moist soils. This is attributed to the rapid increase in the thermal conductivity of coarse grain soils at low saturations. Ochsner et al. (2001) attributed the strong correlation between air volume and thermal conductivity to the low thermal conductivity of air compared to the thermal conductivity of water or soil solid minerals. Air has a thermal conductivity 25 times lower than water and approximately 100 times lower than most soil minerals. Any change of air volume in the sample will also affect the interfacial conduction characteristics of the soil. A reduction in the volume of air due to an increase in volume of solids will cause greater packing of the soil particles and hence lead to better interfacial conduction characteristics. Likewise, a reduction in air due to an increase in water will enhance the inter particle conduction characteristics due to additional moisture at the particle contact points. 3.2. Influence of saturation The effect of increasing saturation on thermal conductivity was determined for each soil and is illustrated in Fig. 5 where the soil thermal conductivity is plotted against saturation at a density of 1500 kg/cm3. As not all experimental tests were conducted at a density of 1500 kg/m3 data points were extrapolated from Figs. 2 and 3 where necessary.

Table 2 Mineralogical composition of soils. Sample

Quartz

Orthoclase/microcline

Albite

Kaolin

Smectite

Mica/illite

Pyrite

Goethite

Residual Siltstone Basaltic Clay Fishermens Bend Silt Coode Island Silt Brighton Group Sand Clayey sand

50 57 51 66

– 3 5 4

3 2 7 3

18 3 6 16

10 35 25 2

19 – 6 6

– – – 3

– – – –

92 63

– –

– –

8 33

– –

– –

– –

– 4

134


3.5

(a)

3.0

Thermal conductivity (W/mK)


3.5

w (%)

2.5

0% 1.2% 3.6% 16.5%

2.0 1.5 1.0 0.5 0.0 1000

1200

1400

1600

1800

2000

(b)

3.0 w (%)

0.1% 5% 11.3% 18.4%

2.5 2.0 1.5 1.0 0.5 0.0 1000

2200

1200

Dry density (kg/m3)

1400

1600

1800

2000

2200

Dry density (kg/m3)

Fig. 2. Thermal conductivity against dry density at a range of moisture contents for (a) Brighton Group sand and (b) Brighton Group clayey sand.

The thermal conductivity of granular soils was observed to increase rapidly at low saturations and then only slightly thereafter. On the other hand, the thermal conductivity of fine grain soils increased at a relatively uniform rate with increasing saturation. The difference in the response of the soils to the increasing saturation is due to the particle size of the soils. Granular soils contain less particle contact points

2.0

(a) Thermal conductivity (W/mK)

Thermal conductivity W/mK

2.0

and hence only a small addition of water is needed to saturate the inter-particle contact points. This significantly increases the thermal conductivity since the thermal conductivity of water is 25 times larger than air. Because the thermal conductivity of the soil solid particles is larger than both air and water the preferred method of thermal transfer is through the soil particles. Hence, any further addition of moisture

1.6

1.2 w (%)

0.8

w (%)

3.7% 13.0% 20.5% 32.0%

0.4

0.0 1000

1200

1400

1600

1800

(b)

1.6

1.2

0.8

w (%)

1.0% 10.0% 19.0% 25.5%

0.4

0.0 1000

2000

1200

Dry density (kg/m3) 2.0

(c) Thermal conductivity (W/mK)


2.0

1.6

1.2

0.8

w (%)

1.2% 10.0% 15.4% 26.4%

0.4

0.0 1000

1200

1400

1600

Dry density (kg/m3)

1400

1600

1800

2000

Dry density (kg/m3)

1800

2000

(d)

1.6

1.2

0.8 w (%)

0.0% 10.7% 19.5% 30.0%

0.4

0.0 1000

1200

1400

1600

1800

2000

Dry density (kg/m3)

Fig. 3. Thermal conductivity against dry density at a range of moisture contents for (a) Basaltic Clay, (b) Fishermens Bend Silt, (c) Residual Siltstone, and (d) Coode Island Silt.


Moist coarse soils Dry coarse soils Fine grain soils

3.0 2.5 2.0 1.5 1.0 0.5 0.0 0.0

0.2

0.4

0.6

3.5

Thermal conductivity (W/m.K)

Thermal conductivity (W/m.K)

3.5

135

Moist coarse soils Dry coarse soils Fine grain soils

3.0 2.5 2.0 1.5 1.0 0.5 0.0 0.0

0.8

0.2

Volume air m3/m3

0.4

0.6

0.8

Volume solids m3/m3

Fig. 4. Thermal conductivity against (a) volume fraction of air and (b) volume fraction of soil solids.

replaces the air between the particles and only slightly increases the thermal conductivity of the granular soils. The steady increase in thermal conductivity with increasing saturation in the fine grain soils is because there are significantly more particle contact points. In fine grain soils saturation must be higher to saturate all the inter-particle contact points.

content of about 20%. It can be noted that higher quartz content results in higher thermal conductivity. Fine grain soils (RS, BC, FBS and CIS) showed significantly lower thermal conductivities than the coarse grain soils (BGS and BGCS) due to their lower quartz content.

3.3. Influence of mineralogy

The thermal conductivity of 54 samples of siltstone, sandstone, and basalt was tested under both water saturated and air saturated conditions using the divided bar apparatus described in Section 3.2. The rocks were collected from various sites around Melbourne between depths of 5 m and 40 m below ground surface. Samples were visually classified as either sandstone or siltstone. The siltstone samples ranged from highly weathered to slightly weathered, the sandstone from highly weathered to moderately weathered and the basalt from extremely to slightly weathered. The core samples were trimmed to approximately 20 mm in length and 50 mm in diameter. Prior to thermal conductivity testing each sample was saturated in a desiccator. The desiccator, an air tight container, was half filled with water and vacuum was applied to remove any

Mineralogy can affect the thermal conductivity of a soil due to different soil minerals having different thermal conductivities. Of particular interest is often the quartz content, as the thermal conductivity of the quartz mineral is 7.7 W/mK compared with most other soil minerals which are typically between 1.8 and 2.8 W/mK (Horai, 1971). In this study the soils contained between 50% and 92% quartz (Table 1) with fine grain soils having lower percentage of quartz content compared to coarse grain soils. However, the majority of the quartz in the fine grain soils was silt or clay sized particles. Fig. 6 presents the variation of thermal conductivity versus quartz content for all soil types with the same density of about 1600 kg/m3 and the same water

4. Thermal conductivity of rocks


2.0

1.6

1.2

0.8

0.4

0.0 0

10

20

30

40

50

60

70

80

Degree of saturation (%) Coode

Residual Siltstone

FBS

Basaltic Clay

BGCS

BGS

Fig. 5. Thermal conductivity against degree of saturation.

90

100

136



2.5

1.0

saturated samples showed a general linear increase in thermal conductivity as the density of the samples increased. This is attributed to the rock particles having a much higher thermal conductivity than air. Rock particles typically have thermal conductivities between 2 and 7.7 W/mK compared to air which is 0.024 W/mK. When the air was replaced with water the thermal conductivity of the rock significantly increased due to the higher thermal conductivity of the water (0.6 W/mK) allowing greater heat flow. Therefore, the influence of density on the thermal conductivity of water saturated samples is significantly less. The results from this study demonstrate no significant relationship between the thermal conductivity of water saturated samples and dry density for the siltstone and sandstone. However, a linear increase in thermal conductivity with density was observed in the saturated basalt samples. The thermal conductivity of the siltstone and to a lesser extent the sandstone was observed to vary considerably between samples of the same or similar densities. This suggests that factors other than density are influencing the thermal conductivity of the rocks. The samples in this study were collected from a number of different sites, boreholes and depths and therefore each sample is unique and may contain slightly different mineralogy and structure. Jones (2003) and Popov et al. (2003) noticed large variations in experimental thermal conductivities between samples of the same rock type. Jones found that mineralogical composition was the main factor influencing the thermal conductivity of the different rocks he tested. Jones also found that other factors such as anisotropy, porosity, and temperature can be important. Popov et al. (2003) tested the thermal conductivity of a number of sedimentary rocks. They found that the anisotropy of the rocks significantly affected their thermal conductivity.

0.5

4.2. Influence of anisotropy

2.3 2.1 1.9 1.7

RS

1.5

BC

1.3

FBS

1.1

CIS

0.9

BGS

0.7

BGCS

0.5 40

50

60

70

80

90

100

Quartz content (%) Fig. 6. Thermal conductivity variation with quartz content for soils.


3.0 2.5 2.0 1.5

0.0 1700

1900

2100

2300

2500

2700

Dry density (kg/m3) Basalt - Sat Sandstone Dry

Basalt - Dry Siltstone - Sat

Sandstone - Sat Siltstone - Dry

Fig. 7. Thermal conductivity against dry density for rock samples.

dissolved air while the rock samples submerged in the water were getting saturated. The samples were weighed every few days and assumed to be fully saturated when the weight stopped changing. Once the samples were fully saturated the saturated thermal conductivity was tested in the divided bar apparatus. The samples were then oven dried at 105 °C and re-tested to measure the dry thermal conductivity. A summary of the experimental thermal conductivity results is presented in Fig. 7 and Table 3 for both air and water saturated samples.

4.1. Influence of density The relationship between thermal conductivity and density can be observed for both air and water saturated rocks in Fig. 7. The air

The anisotropy of rocks has been found by many researchers to influence the thermal conductivity of sedimentary rocks (Midtomme and Roaldset, 1999; Beardsmore and Cull, 2001; Popov et al., 2003). Midttomme and Roaldset (1999) reported thermal conductivities parallel to the grain were in some cases up to twice that measured perpendicular to it in sedimentary rocks. While Beardsmore and Cull (2001) noticed in several documented cases that the thermal conductivity of shale does not increase with density the same way that other rocks do. This is because shale is comprised of highly anisotropic silicate sheets. The thermal conductivity parallel to the mineral sheet is often many times greater than perpendicular to it and as the density of the shale increases the particles arrange themselves into a preferred horizontal alignment. In this study the bedding orientation was used to investigate the influence of anisotropy of siltstone samples. The bedding orientation was measured for 30 of the 35 siltstone samples and plotted against the thermal conductivity (Fig. 8). A bedding orientation could not be definitively recognized in five of the siltstone samples and these samples were subsequently excluded from Fig. 8. In general the figure illustrates that the thermal conductivity increased with increasing bedding orientation. This highlights the anisotropic nature of the siltstone samples tested in this study.

Table 3 Summary of rock thermal conductivities of rocks from this study. N

Dry density (kg/m3)

Ave. dry density (kg/m3)

35 10 9

2130–2490 2030–2470 1750–2610

2320 2250 2290


Ave. thermal conductivity (W/mK)

Saturated Siltstone Sandstone Basalt

1.73–2.47 2.46–2.84 1.08–1.57


Ave. thermal conductivity (W/mK)

Dry 2.14 2.64 1.39

0.72–1.54 1.04–1.67 0.56–1.05

1.23 1.3 0.88


137


3.0 2.5

Heat Flow

2.0 Bedding 0˚

1.5 Heat Flow

1.0 0.5 Bedding 90˚

0.0

0

20

40

60

80

Bedding orientation (degrees) Fig. 8. Thermal conductivity against bedding orientation for siltstone samples.

4.3. Influence of mineralogy In this study rock samples were collected from different sites, boreholes and depths. Considering this, it is likely that each sample contained a slightly different mineralogical makeup. As discussed in Section 3.3, quartz is the main mineral that influences the thermal conductivity of soils and rocks and in this study the percentage of quartz was used to assess the variation in the thermal conductivity of the rocks. The quartz content of ten samples of siltstone were analysed in this study. Samples were strategically selected to cover a range of densities, thermal conductivities and bedding orientations. The samples were sent to the University of Ballarat where thirty micron, thin sections were produced for each sample. Petrographic analysis was performed on each of the thin sections and the quartz content quantitatively calculated. The results of the petrographic analysis yielded quartz contents of the siltstone between 10% and 26%. Fig. 9 presents the quartz content plotted against the thermal conductivity. Of the ten samples, three samples had bedding orientations of 17° and densities of approximately 2350 kg/m3 (solid circles) and two samples had bedding orientations of 35° and densities of approximately 2350 kg/m3 (solid triangle). The

5. Conclusions

1.5

Soil and rock thermal conductivity is a key parameter in the design of borehole heat exchangers and geothermal energy foundations. This study has presented a comprehensive laboratory analysis of the thermal conductivity of soils and rocks over a wide range of moisture contents and densities. The soils tested showed that thermal conductivity varied with soil moisture content, dry density, mineralogical composition and particle size. The thermal conductivity of the soils was observed to increase with an increase in dry density and moisture content. The thermal conductivity of the rocks was observed to increase with increasing density when tested dry. However, when the rock samples were water saturated only the basalt rocks showed a correlation between density and thermal conductivity. The siltstone and sandstone rocks showed no distinct correlation between density and thermal conductivity. The difference between the thermal conductivity of samples of saturated siltstone and sandstone rocks were attributed to a combination of different mineralogical composition, anisotropy and density. The soil and rock data collated in this study provides new knowledge of the thermal properties of soils and rocks from the Melbourne region. This knowledge provides an invaluable resource for the estimation of thermal conductivity for the design of thermally active ground structures built in Melbourne. Properties such as dry density, moisture content and mineralogical composition need to be taken into account when assessing the thermal conductivity of soils/rocks below ground surface for the design of such structures.

1.0

Acknowledgements

3.0


remaining samples (crosses) contained a range of bedding orientations and densities. The thermal conductivity was observed to generally increase with increasing quartz content for all samples. The variation observed between samples is attributed to the different bedding orientation and density.

2.5

2.0

This project was funded by the Victorian Government Sustainability Fund, Golder Associates Pty. Ltd., Vibropile Pty. Ltd., Geoexchange Australia Pty. Ltd. and Genesis Now. Their support is gratefully acknowledged.

0.5

0.0

0

5

10

15

20

25

Quartz content (%) Bedding 17°

Bedding 35°

Bedding 10° - 70°

Fig. 9. Thermal conductivity against quartz content for ten siltstone samples.

30 References Abuel-Naga, H.M., Bergado, D.T., Bouazza, A., 2008. Thermal conductivity evolution of saturated clay under consolidation process. International Journal of Geomechanics 8 (2), 114–118.

138


Abuel-Naga, H.M., Bergado, D.T., Bouazza, A., 2009. Laboratory and field thermal conductivity of a saturated clay. Engineering Geology 105 (3–4), 211–219. Abu-Hamdeh, N.H., Reeder, R.C., 2000. Soil thermal conductivity: effects of density, moisture, salt concentration, and organic matter. Soil Science Society American Journal 64, 1285–1290. Beardsmore, G.R., Cull, J.P., 2001. Crustal Heat Flow: A Guide to Measurement and Modelling. Cambridge University Press. Bouazza, A., Singh, R.M., Wang, B., Barry-Macaulay, D., Haberfield, C., Chapman, G., Baycan, S., Carden, Y., 2011. Harnessing on site renewable energy through pile foundations. Australian Geomechanics 46 (4), 79–90. Brandon, T.L., Mitchell, J.K., 1989. Factors influencing thermal resistivity of sands. Journal of Geotechnical Engineering 115 (12), 1683–1698. Carslaw, H.S., Jaeger, J.C., 1959. Conduction of Heat in Solids, 2nd edition. Oxford, London. Chen, S.X., 2007. Thermal conductivity of sands. Heat and Mass Transfer 44 (10), 1241–1246. Dali Naidu, A., Singh, D.N., 2004. A generalized procedure for determining thermal resistivity of soils. International Journal of Thermal Sciences 43 (1), 43–51. DeMoel, M., Bach, P., Bouazza, A., Singh, R.M., Sun, O., 2010. Technological advances and applications of geothermal energy pile foundations and their feasibility in Australia. Renewable and Sustainable Energy Reviews 14 (9), 2683–2696. Decagon Devices, 2006. KD2 Pro Thermal Properties Analyzer Operator's Manual. Decagon Devices, Inc. Ewen, J., Thomas, H.R., 1987. The thermal probe: a new method and its use on an unsaturated sand. Geotechnique 37 (1), 91–105. Farouki, O., 1986. Thermal Properties of Soils, vol. 11. Trans Tech Publications, Clausthal-Zellerfeld, Germany. Horai, K., 1971. Thermal conductivity of rock-forming minerals. Journal of Geophysical Research 76 (5), 1278–1308. Jackson, R.D., Taylor, S.A., 1986. Thermal conductivity and diffusivity. Methods of Soil Analysis, Part 1. Physical and Mineralogical Methods, pp. 945–956. Johnston, I.W., Narsilio, G.A., Colls, S., 2011. Emerging geothermal energy technologies. KSCE Journal of Civil Engineering 15 (4), 643–653.

Jones, M.Q.W., 2003. Thermal properties of stratified rocks from Witwatersrand gold mining areas. Journal of South African Institute of Mining and Metallurgy 103 (3), 173–186. Mickley, A.S., 1951. Thermal conductivity of moist soil. American Institute of Electrical Engineers Transactions 70 (2), 1789–1797. Midttømme, K., Roaldset, E., 1999. Thermal conductivity of sedimentary rocks; uncertainties in measurement and modelling. Geological Society, London, Special Publications 158, 45–60. Mitchell, J.K., Kao, T.C., 1978. Measurement of soil thermal resistivity. Journal of the Geotechnical Engineering Division, ASCE 104 (GT10), 1307–1320. Nakshabandi, G., Kohnke, H., 1965. Thermal conductivity and diffusivity of soils as related to moisture tension and other physical properties. Agricultural Meteorology 2, 271–279. Ochsner, T.E., Horton, R., Renen, T., 2001. A new perspective on soil thermal properties. Soil Science Society of America Journal 65 (6), 1641–1647. Penner, E., Johnston, G.H., Goodrich, L.E., 1975. Thermal conductivity laboratory studies of some MacKenzie highway soils. Canadian Geotechnical Journal 12 (3), 271–288. Popov, Y., Tertychnyi, V., Romushkevich, R., Korobkov, D., Pohl, J., 2003. Interrelations between thermal conductivity and other physical properties of rocks: experimental data. Pure and Applied Geophysics 160 (5–6), 1137–1161. Sass, J.H., Stone, C., Munroe, R.J., 1984. Thermal conductivity determinations on solid rock — a comparison between a steady-state divided-bar apparatus and a commercial transient line-source device. Journal of Volcanology and Geothermal Research 20, 145–153. Singh, R.M., Bouazza, A., 2013. Thermal conductivity of geosynthetics. Geotextiles and Geomembranes. http://dx.doi.org/10.1016/j.geotexmem.2013.06.002. Van Rooyen, M., Winterkorn, H.F., 1957. Structural and textural influences on thermal conductivity of soils. Highway Research Board Proceedings. , 39, pp. 576–621. Wang, B., Bouazza, A., Barry-Macaulay, D., Singh, R.M., Haberfield, C., Chapman, G., Baycan, S., 2012. Field and laboratory investigation of a heat exchanger pile. State of the Art and Practice in Geotechnical Engineering, 225. ASCE Geotechnical Special Publication, San Francisco, pp. 4396–4405.