2024-Soil Thermal

Soil Thermal
Properties (STP) for

subsea cable design
Cables technical working group
March 2024
SOIL THERMAL PROPERTIES FOR SUBSEA CABLE DESIGN
BEST PRACTICE GUIDE
Document information
OWA research area: Cables technical working group (TWG-C)
Project title: Soil Thermal Properties (STP) for Subsea Cable Design
Contractor: Cathie, ITP Energised, and University of Southampton
Carbon Trust contact email: karolina.zieba@carbontrust.com
OWA parties
Participants in the Offshore Wind Accelerator (OWA) programme (Stage IV): EnBW Energie Baden-Württemberg AG, RWE
Offshore Wind GmbH, Equinor ASA, Ørsted Wind Power A/S, ScottishPower Renewables (UK) Limited, Shell Global Solutions
International B.V, SSE Renewables Services (UK) Limited, Vattenfall Wind Power Ltd, TotalEnergies OneTech, and the Carbon
Trust.
Document history
Version Date Main modification Lead Author
1.0 02/05/2023 Draft CATHIE / UoS / ITPE
2.0 16/06/2023 For comment CATHIE / UoS / ITPE
3.0 15/12/2023 For approval CATHIE / UoS / ITPE
Delivery partners:
Disclaimer
This report is issued by the Carbon Trust on behalf of the Offshore Wind Accelerator (“OWA”). While reasonable steps have
been taken to ensure that the information contained within this report is accurate, the authors, the Carbon Trust, its agents and
consultants and the partners and developers within the OWA (and each of them), to the fullest extent permitted by law, shall
not have nor be deemed to have (1) a duty of care to readers and/or users of this report, (2) made or given or to make or give
any warranty or representation (in each case whether express or implied) as to its accuracy, applicability or completeness
and/or (3) or have accepted any liability whatsoever for any errors or omissions (whether negligent or otherwise) within it. It
should also be noted that this report has been produced from information relating to dates and periods referred to in it. Users
and readers use this report on the basis that they do so at their own risk. The intellectual property rights in this report shall be
deemed, as between readers and users of this report and the Carbon Trust, to belong to the Carbon Trust.
© The Carbon Trust 2024

BEST PRACTICE GUIDE
Contents
1. Background 1
1.1 Introduction 1
1.2 Impacts of the thermal environment on cable rating and cable cost 2
1.3 Characterising the seabed thermal environment for cable design 5
2. Thermal soil property collation and measurement 7

2.1 Thermal properties within the site investigation process 7
2.2 Methods for measuring thermal conductivity 7
2.3 Laboratory testing techniques 8
2.4 Sample preparation 9
2.5 In situ measurement techniques 10
3. Selection of soil thermal properties 13

3.1 Introduction 13
3.2 Thermal property geodatabases 14
3.2.1 Statistical analysis 14
3.2.2 Repeatability-correction of TC database for each soil type 16
3.2.3 Weighted database according to seafloor soil variation 16
3.2.4 Thermal conductivity versus porosity – trend lines per soil type 17
3.3 Pre survey estimation from databases 18
3.3.1 No site specific data available (Data Scenario 0) 18
3.3.2 Soil type data available (Data Scenario 1) 19
3.4 Site specific data available excluding thermal properties 19
3.4.1 In situ CPT data available (Data Scenario 2) 19
3.4.2 Lab test data available (Data Scenario 3a) 19
3.5 Site specific thermal property data available (Data Scenario 3b) 20
3.6 Modelling trench effects – allowing for soil disturbance (Data Scenario 4) 20
BEST PRACTICE GUIDE
List of tables
Table 1-1: Data scenarios for Thermal Conductivity estimation 6
Table 2-1: Comparison of laboratory test methods 8
Table 2-2: Laboratory-based versus in situ geotechnical tests 12
Table 3-1: Data scenarios for Thermal Conductivity estimation (Table 1-1 repeated) 13
Table 3-2: Normal distribution statistical parameters – geodatabase evaluation 16
Table 3-3: Relevant phase relationships to derive porosity from standard laboratory data 17
Table 3-4: Typical grain thermal conductivities of marine sediments 18
List of figures
Figure 1-1: Steady state current vs thermal resistivity – 1200 & 1600 mm2 cables 3
Figure 1-2 Schematic representation of cable CSA (vs base case scenario) - steady state 3
Figure 1-3: Continuous static cable rating for a generic AC cable 4
Figure 1-4: Illustration of data scenarios 5
Figure 2-1: Common methods for evaluating thermal conductivity of marine sediments 7
Figure 3-1: Illustration of data scenarios (Figure 1-4 repeated) 13
Figure 3-2: Combined box, whisker and beeswarm scatter plot - statistical parameters 15
Figure 3-3: Cable temperature change due to burial-induced seabed properties changes 21
List of appendices
- Techniques for measuring thermal conductivity
Appendix A.1 Laboratory techniques
Appendix A.1.1 Needle probe
Appendix A.1.2 Transient plane source (hot disk)
Appendix A.1.3 Guarded hot plate
Appendix A.1.4 Thermal cell
Appendix A.2 In situ techniques
Appendix A.2.1 Passive and active thermal probes
Appendix A.2.2 Passive method – thermal CPT
Appendix A.2.3 Passive method – correlation with conventional CPT measurements
- References
BEST PRACTICE GUIDE
1. Background
1.1 Introduction
The dissipation of thermal energy from subsea HV cables is a key consideration for:
(i) the design and manufacture of subsea cables;
(ii) dynamic rating of the cable during the lifetime of operation;
(iii) the interpretation of distributed temperature sensor (DTS) data for monitoring cable
performance, depth of burial, transgression of the “2K criteria” of temperature rise to protect benthic
life, and finally, in extreme circumstances, fault location.
This study demonstrates that a proper consideration of the thermal environment, using geotechnical
best practices, could save tens of millions of pounds of capital expenditure, whilst also helping to
reduce operational costs. For instance, offshore cables represent 30% of offshore wind farm insurance
claims incurred and 50% of the total value of claims between 2010 and 2020 (GCube, 2021).
The environmental factors that control heat dissipation from subsea HV cables are the thermal
conductivity, volumetric heat capacity and permeability of the seabed soils as well as the lifetime
changes in the depth of cover over the cable. The rate and magnitude of heat dissipation, conductor
temperature and ultimately the maximum power transfer possible can be determined by the IEC
standards or derived from the fundamental physical laws solved analytically or using numerical
methods. The factor that has the greatest influence on heat dissipation is the thermal conductivity in
most submarine cable environments. Hughes et al (2015) have demonstrated that the permeability of
soils can also impact the heat dissipation by enabling convective heat transfer. However, this only
becomes an issue for the higher permeabilities found in unimodal medium to coarse sands and gravels
and is therefore not considered in detail within these guidance notes.
With this background, the key objective of this Best Practice Guide is to:
• Provide an understanding of the impact of thermal conductivity on cable design and identify
the influence of the accuracy and precision of these measurements on the cable design and
operating condition;
• Present appropriate engineering practices to support measurement of the thermal
conductivity of marine soils, highlighting the different levels of accuracy and precision that
can be achieved by different methods;
• Provide an understanding of the potential changes of these thermal properties in response to
installation effects – from trenching or ploughing and cable burial – and the implications of
these changes in soil properties for cable design and operation.
These guidelines are relevant at all of the phases of the OWF including design, installation and
operation. The guidelines are a deliverable from the Carbon Trust project “Soil Thermal Properties for
Subsea Cable Design”. They have been developed from three detailed reports produced earlier in the
project which covered “Cost Implications of Thermal Conductivity Values” (WP1), a “Testing
Methodology Review” (WP2) and “Trenching Effects on the Design Parameter Framework” (WP3).
Selected sections from those reports are provided in Appendix A in this guideline and are referenced
within the main text.
1
BEST PRACTICE GUIDE
Thermal conductivity, in W/mK, is the key parameter measured or reported during geotechnical testing.
However, its reciprocal, thermal resistivity (Km/W), is the form used in all electrical cable design and
post installation monitoring software and so both will be referred to within this document. These
guidance notes are also solely focused on soils in marine conditions, which results in them being fully
and permanently saturated. Landfall sites, which experience spatially and temporally varying saturation
states, create a significantly more complex thermal environment and therefore require a different
thermal assessment approach. This guidance is not appropriate for unsaturated soils.
1.2 Impacts of the thermal environment on cable rating and cable cost
WP1 of this project provided a series of Case Studies designed to illustrate the implications of soil
thermal properties on subsea cable design, with specific emphasis on the cable cross sectional areas
(CSA’s) required to achieve a set rating (i.e. maximum current) and the subsequent cost implications of
these choices. Each study is compared against a base case of a generic three phase, AC, 1400mm2
CSA, export cable buried at 1 m depth within a seabed with a fixed thermal resistivity of 0.7 Km/W
(which is the standard value quoted in the international standards, IEC-60287 and 60853, and previously
used extensively across the sector). The base case has a rating of 1000A. These Case Studies
demonstrate the impacts of the following variables on cable design:
• Soil type;
• Uncertainty of thermal measurements;
• Choice of statistical parameter (i.e. fractile within scattered data) for unitised thermal
resistivities;
• Impact of trenching on the thermal environment.
Figure 1-1 demonstrates that for two different sizes of cable (1200 and 1600mm2 respectively) the
cable rating can vary by between 300-350A (~35%) depending on both the thermal properties of the soil
and which thermal resistivity value (Maximum, P90, P75 and Mean) is used to represent the
environmental conditions (WP1). Such a variation is financially and operationally important and is
comparable to similar ratings improvements in terrestrial environments that have been considered
significant in other studies (e.g. Pilgrim et al, 2012).
2
BEST PRACTICE GUIDE
Figure 1-1: Steady state current vs thermal resistivity – 1200 & 1600mm2 cables
Note: Plotted are the values for mean, P75, P90 and Max TR values for each soil type
Similarly, Figure 1-2 demonstrates that a 20% uncertainty in the thermal conductivity measurement can
lead to a cost differential of up to £48.5 million for a 100km cable when compared against the base
case scenario (based on October 2021 materials costs – WP1). Even for a 10% uncertainty in the
thermal conductivity, the maximum rating for a single cable can vary between approximately ±1.5%
(high thermal conductivities more than ~2W/mK) and ~±5% (low thermal conductivities [less than
~1W/mK]). These impacts increase with increasing burial depth (Figure 1-3). It is therefore essential to
adopt the most precise measurement methods possible, for both laboratory based and in situ thermal
conductivity measurements and understand the uncertainty budget of all methods used.
Figure 1-2 Schematic representation of cable CSA (vs base case scenario) - steady state
Note: including ±20% uncertainty error in thermal resistivity for a range soil types
3
BEST PRACTICE GUIDE
Figure 1-3: Continuous static cable rating for a generic AC cable
Notes: Rating satisfying the maximum permissible conductor temperature of 90°C, for a range of varying
thermal conductivities and burial depths
These observations are based on using thermal conductivity measurements of the pre trenched, ambient
soil condition, which is the condition in which soil is tested in the laboratory and in situ. However, it is
important to consider how the key thermal properties may change during and after the installation
process. There are no publicly available direct measurements of thermal conductivity of trenched soils.
As an alternative, WP3 modelled the potential impact of trenching on thermal conductivity by estimating
the change in thermal properties via extant relationships between relative density [for which pre- and
post-trenching data is available] => void ratio / porosity => thermal conductivity. These suggest that:
• Jetting and ploughing of sands result in an increase in void ratio and therefore in conductor
temperature with 75% of the modelled scenarios resulting in an increase of more than 5°C.
However, these relate to the largest inferred changes in relative densities that occur
immediately after installation. It is probable that trench conditions will undergo at least some
post installation densification and therefore thermal conductivities will trend back towards the
pre installation conditions and hence a smaller increase in conductor temperature;
• Jetting and ploughing in soft clays also results in a relative increase in conductor temperature
relative to the non-trenched condition but at a level (less than 3°C) that is not significant for
cable design as it is within the uncertainties of the cable component of all IEC models;
• Jetting in stiff clays, which frequently results in a blocky backfill within the trench, led to
significant decreases (between 5°C and 33°C) or increases (between 1°C and 7°C) in
temperature depending on a combination of factors: (i) the block to matrix volume ratio, (ii)
the block distribution or packing arrangements, and most importantly (iii) the nature of the
fluid or consolidated soil material within the matrix;
• A cut trench in stiff clay that naturally backfills with marine sand has contrasting thermal
conductivities between the in situ clay material and the backfill. In addition, the fill has higher
4
BEST PRACTICE GUIDE
permeability. Together these factors result in a significant temperature drop of between ~7°C
and 11°C. This is due to the enhanced conduction and convection through the sand backfill.
The idealised nature of the models examined means the relevant results should not be used as
quantitative values of change in conductor temperature (and therefore cable rating) for any specific
project site. Instead, the aim is to highlight the need to accurately characterise soil thermal properties to
avoid over- or under-estimating the heat diffusion from the cable, and therefore potentially overheating
the cable and soil or overdesigning the cable.
In addition, where these case studies involve assumptions about the final state of the ground after
trenching, e.g. the form of any backfill blocks and matrix infill, they highlight the potential value of in situ
testing in and around installed cables or analysis of DTS data to help resolve these gaps in knowledge.
The financial gains of optimised cable design through an accurate knowledge of the seabed thermal
environment could be further enhanced through understanding the full financial impacts of:
• Micro routing to optimise the thermal environment;
• Including thermal considerations in the choice of trenching technique and potentially
proactively pre conditioning soils and cable attachments to enhance the thermal diffusion;
• Enhancing operational condition monitoring through the confident use of Real Time Thermal
Rating (RTTR) and “Depth of Burial” software thus optimising power transmission and allowing
a reduction in the number of post installation geophysical and visual monitoring surveys.
1.3 Characterising the seabed thermal environment for cable design

To support optimised cable thermal design, through accurate characterisation of the seabed thermal
properties, a sequence of idealised “data scenarios” is illustrated in Figure 1-4 and summarised in Table
1-1. These staged scenarios broadly represent the evolution of data availability during a project. They
represent the available information about the soil conditions and the soil thermal environment and
illustrate the progression of this knowledge during the different phases of site investigation.
Figure 1-4: Illustration of data scenarios

Additional data sets become available during project development
Data
Scenarios
Thermal
conductivity
range
5
BEST PRACTICE GUIDE
Table 1-1: Data scenarios for Thermal Conductivity estimation

Data
Basis for estimating
scenario / Data type Key data Section
Thermal Conductivity
survey stage
No site specific Full thermal database

0 None Section 3.3.1
data scaled for region
1 Geological setting Soil type Soil type database subset Section 3.3.2
Direct Thermal Conductivity

T-CPT measurement measurement or Sections 2.5 and
2 In situ testing
or CPT profile Correlated porosity & 3.4.1
database
Moisture content, Correlated soil type,

3a Basic lab testing Section 2.3 and 3.4.2
specific gravity porosity & database
Thermal
Measurement of Measured soil type,
3b Conductivity lab Section 2.3 and 3.5
Thermal Conductivity porosity & database
test
Modelled Effect of PSD – to estimate

Changed porosity (using Id)
4 trench on Thermal change in porosity Section 3.6
& isomaterial line
Conductivity from in situ value
This document provides guidance on the measurement of thermal properties and the selection of design
values for cable design and condition monitoring software. The guidance is divided as follows:
• Section 2: Guidance on options for measuring the thermal properties of marine soils.
• Section 3: Guidance on the selection of design thermal properties, understanding uncertainty,
when equipped with different types of data (representing typical different survey stages). This
guidance includes approaches for adjusting thermal properties to allow for soil disturbance
effects associated with the trenching and burial process of cables.
In addition, Appendix A complements the guidelines by discussing currently available techniques for
measuring thermal conductivity.
6
BEST PRACTICE GUIDE
2. Thermal soil property collation and measurement
2.1 Thermal properties within the site investigation process

To properly characterise the soil thermal environment along a cable route section, a sufficient number of
thermal conductivity/resistivity measurements should be performed to fully characterise the near seabed
geology and the requirements for cable burial and cable protection. It is recommended that both
preliminary (typically based only on geophysical survey data and desk study geotechnical data) and final
(integrated, site specific geophysical and geotechnical data) ground models are consulted to inform the
location and frequency of sampling. Thermal measurements should extend to below the target burial
depth. Materials that potentially limit the thermal diffusion (e.g. organic or carbonate rich soils) should
be specifically targeted.
Due to the variability in the length and criticality of cables, and the variability in the degree of soil
heterogeneity, no ideal number or spacing of tests can be provided. Sufficient measurements should be
made to provide statistically significant sets that support mean and low/high estimates of thermal
properties for each distinct soil unit. These typically require at least several hundreds to 1000+ tests per
OWF, also depending on the cable route length and encountered soil conditions.
2.2 Methods for measuring thermal conductivity

The most common approach in site investigation studies has been to measure thermal conductivity on
retrieved soil samples in the laboratory using a range of both transient and steady state methods. This
approach requires sampling cores from the seabed and either measuring directly on the cores offshore
after retrieval, or on samples onshore. The samples tested onshore may be intact or may have been
reconstituted/remoulded from disturbed (e.g. bagged) samples. Intact samples still require handling and
preparation for laboratory testing, and the testing process can also introduce disturbance.
It can therefore be challenging to recreate the in situ ground conditions – i.e. voids ratio/porosity and soil
fabric/structure – in the laboratory and requires sufficient expertise. For this reason, the development of
robust and well validated methodologies for in situ measurement of thermal properties is the focus of
current research and innovation.
Common methods for evaluating thermal conductivity of marine sediments are shown in Figure 2-1.
Figure 2-1: Common methods for evaluating thermal conductivity of marine sediments
7
BEST PRACTICE GUIDE
2.3 Laboratory testing techniques

Laboratory techniques are subdivided into two broad groups:
• Transient methods (needle probe and hot disk) which measure the temperature variation of
the sediment in response to an instantaneous heat pulse;
• Steady state methods (guarded hot plates and thermal cells) that apply a constant heat flux to
the sample until a stable temperature distribution is achieved.
Table 2-1 presents a summary of the advantages and disadvantages for the four primary laboratory
methods used to measure thermal conductivity. A detailed review of each method and the commonly
available systems is provided in Appendix A.
Table 2-1: Comparison of laboratory test methods
Test
Advantages Disadvantages Accuracy and Precision
Method
• Short test duration • Ensuring good ASTM D5344 (2014):

• Suitable for most contact with soil Precision ±10-15% (Repeatability
Needle soils sample and Reproducibility)
Probe • Standardised test • Delicate
method • Not suitable for
coarse soil
• Short test duration • Requires two No standards for soils but BS EN
Transient • Minimal sample identical specimens ISO, 22007-2 (2015) for plastics &
Plane preparation • Limited experience thermal insulation material suggest
Source • Also determines in being used for an accuracy of ±2-5% and a
(Hot Disk) thermal diffusivity & testing soils precision (repeatability) of ±1-2%
heat capacity
• Standardised • Long test duration BS 12664 (2001) – Based on
method • May not be suitable Building Materials
Guarded
• Easy to use for most soils (< Accuracy ±2%
Hot Plate
• Effective for gravels 2W/mK) Precision ±0.5-1% (Repeatability and
• Moisture migration Reproducibility)
• Suitable for most • Long test duration Comparable to Guarded Hot Plate
soils • Moisture migration method
Thermal • Minimal sample • Bespoke equipment BS 12664 (2001)
Cell disturbance Accuracy ±2%
Precision ±0.5-1% (Repeatability and
Reproducibility)
To achieve the highest accuracy and precision of the thermal conductivity measurements the following
requirements should be met:
• Device specification, procedures and calibration steps should meet the relevant standards and
the manufacturer’s requirements;
• The sample should be in equilibrium with the ambient temperature prior to testing;
• All devices should be calibrated with an appropriate medium and the calibration should be
applied to the final results. It is recommended to use Ballotini as an effective calibration
material for marine soils;
8
BEST PRACTICE GUIDE
• A constant and stable power supply should be applied during the test;
• Metadata of all of these aspects should be well recorded and included in the test reporting.
Additional requirements specifically for needle probe tests, which are the most commonly used test:
• The device should be fully embedded and have good contact with the soil sample (for example.
the insertion process should be a clean direct push without lateral movement);
• The device should be installed without hammering or excessive force;
• For multiple measurements, tests should be performed with sufficient distances between
them and from the sample boundary to avoid interaction effects;
• The needle probe should have a diameter greater than the grain size of the soil sample;
• Heating boundary effects should be avoided. This is often approximated by ensuring the
diameter of the soil sample container is 50x the needle diameter, which gives typical sample
diameters of 50-100mm.
• The sensitivity of the calculated thermal conductivity to the choice of the section of the cooling
or warming curve should be quantified and recorded.
In addition, for best practice, it is recommended that:
• The results and interpretation (i.e. the measured test response and the interpretation or
graphical construction used to find thermal properties) should be inspected by the user or
report writer visually to identify possible errors or alternative interpretations;
• The results should be ‘sense checked’ against other available data for similar soil conditions
and properties, to highlight any gross differences and unexpected results, which merit further
investigation;
• Information relevant to the specific soil conditions should be considered prior to testing to
ensure that the test arrangements are appropriate for the soil type;
• The derived thermal properties should be evaluated in combination with the soil
characterisation (for example, by plotting trends of thermal conductivity with moisture content
or voids ratio in each soil unit, to assemble sets of similar results from which statistical
measures of best, low and high thermal conductivity can be identified).
2.4 Sample preparation

Heterogeneity and sample disturbance (either during sampling, transportation or reconstitution in the
laboratory) are considerable sources of variability in thermal conductivity measurements and can often
exceed the uncertainty associated with the measurement and its interpretation. Consequently, it is
suggested that all laboratory samples are prepared with the following considerations:
• Saturation states should be close to 100% and always above 80%, with saturation ideally being
undertaken with a saline solution;
• Where reconstitution/compaction is attempted, the source of the target densities should be
clearly defined, and attempts to recreate the in situ density should avoid creating artificial
levels of anisotropy perpendicular to the applied stress planes, which can occur from tamping;
• Where possible laboratory measurements of density should be compared with measurements
taken on the vessel immediately on retrieval of the specimen. Even though disturbance is
inevitable during the sampling process, water loss is common on retrieval/splitting of the
9
BEST PRACTICE GUIDE
cores. Therefore, measurements of moisture content and/or density from the vessel provide
data closest to the natural condition and therefore a suitable benchmark;
• The sample being prepared should aim to represent the in situ soil structure and fabric in terms
of grain sizes and mineralogy and also small scale heterogeneities such as fine scale bedding.
Consideration should be given to how this structure and fabric could change during the
process of sampling and thermal testing, but also how it may change during cable installation
(see Section 3.6).
2.5 In situ measurement techniques

To make in situ measurements of thermal conductivity there are currently three principal approaches:
• A heat pulse is produced from an active source that is located along a linear sensor string that
is attached to a coring system or solid probe. A range of instruments are available typically
derived from an early “violin bow” design developed for academic research. The temperature
decay curve of the heat pulse is recorded at multiple sensors from which a profile of thermal
conductivity is calculated along the length of the probe;
• A passive source, where the heat pulse is generated through friction between the probe and
the seabed soil during insertion of the probe. This approach has recently been developed as a
modification of a conventional CPT with an added temperature sensor, frequently in the tip,
and known as a T-CPT. As for active sources, the sensor records the rate of temperature decay
to calculate thermal conductivity at a set depth. As these systems only have a single
temperature sensor, multiple pushes are required to get measurements at different depths.
• Conversion from standard CPT parameters (cone resistance and friction ratio) to give bulk
thermal conductivity values through correlations. These correlations usually first estimate the
bulk density, with correlation parameters being linked to grain mineralogy. The voids ratio and
porosity is then estimated from the bulk density, based on an assumed grain specific gravity
linked again to the soil mineralogy. From porosity the bulk thermal conductivity is then
estimated, using assumed values for the pore fluid and grain thermal conductivities (with the
grain thermal conductivity also being estimated from the soil mineralogy).
For successful deployment and the acquisition of reliable measurements it is important to consider the
following, for the active and passive sources:
• The ambient temperature should be well constrained. For active systems a time lag of several
minutes is advised for the frictional heat pulse from installation to dissipate prior to generating
the controlled heat pulse. For passive systems the ambient temperature can either be
backcalculated (but with inevitable uncertainty) or temperature measurements should be
continued until the ambient temperature has been achieved again;
• The specific thermal properties of the probe (i.e. the parts, their shapes and their thermal
properties) and the specific contact thermal resistance with the surrounding soil require all in
situ probe thermal measurements to be calibrated either against theoretical models or
conventional needle probe measurements either in the field or the laboratory. The contact
thermal resistance can be affected by liquefaction of the surrounding soil during installation,
convection in the pore water and disturbance of the adjacent grains resulting in a local change
in porosity. As discussed in Sections 2.3 and 2.4, for in situ tests to be calibrated against
needle probe measurements, they should themselves be calibrated against appropriate
10
BEST PRACTICE GUIDE
standards. Whatever approach to calibration is used it needs to be clearly recorded and set
out in the test report;
• A measurement time of at least 500 – 1200 seconds is required to record the full response.
For passive systems it is also important to consider that:
• The initial temperature rise ∆T should exceed 3°C, to provide a clear decay curve. This is
usually achieved in sands where probe shaft friction is high, but is difficult in fine grained soil.
There are still few publicly available studies that have tested the accuracy and precision of all of the
active and passive in situ techniques. Published results indicate that the “relative accuracies” of active
systems are within ±5% of laboratory thermal conductivity measurements for the same material.
Unpublished commercial studies have shown thermal conductivity measurements in sands using an in
situ passive system that are 20% higher than measurements from needle probe tests on collocated
samples. However, as further experience of the passive in situ CPT systems is gained, this observation
may not be universal. There is therefore an urgent requirement for such systems to undergo well
documented benchmark testing, comparable to that undertaken for the ASTM D5344 needle probe
standards.
The CPT correlation approach to estimate thermal conductivity from standard CPT data are attractive
due to the simplicity and potential time and cost savings that arise from eliminating any specialised
thermal testing. However, at present there are quite large differences both when compared against
laboratory needle probe measurements and collocated in situ T-CPT and certainly at levels that could
significantly impact on cable design as described in Section 1.2. Consequently, this approach should at
present be treated with caution until a realistic accuracy and precision can be determined. It will remain
a useful tool to gain estimates of thermal conductivity from early site investigation campaigns where
thermal test data does not yet exist, or to supplement actual measurements across a project or soil unit
after the correlation has been checked and/or recalibrated against either laboratory or other in situ
measurement methods.
A summary of the relative merits of laboratory based and in situ thermal property tests is presented in
Table 2-2. This comparison highlights that several factors influence the optimum choice of test and there
is no single test type that is preferred.
Our expectation is that in situ testing will become the default approach to measuring thermal conductivity
for cable design, with laboratory based measurements being used to further analyse complicated and
thermally limiting soils (e.g. gravels, very fine grained over consolidated materials, organic and carbonate
rich materials). However, since active in situ testing apparatus is not widely specified for site
investigations and passive in situ tests are relatively new, with limited published experience, in situ testing
is not yet accepted as the default approach, and best practice is currently to use both types of testing.
Needle probe laboratory testing of tube samples remains currently the primary source of thermal
conductivity data on projects.
11
BEST PRACTICE GUIDE
Table 2-2: Laboratory based versus in situ geotechnical tests
Laboratory based measurement In situ measurement

Easily control both environmental,
Control of Experimental environmental, testing
testing and boundary conditions. NB.
experimental and boundary conditions difficult if not
Boundary conditions can be restricted
conditions impossible to control
depending on the equipment
Generally yes as the sample is logged Interpretation is difficult without
Interpretability
i.e. well characterised adjacent laboratory soil samples
Can be restrictive to unrepresentative,
Generally larger volumes are sampled
Representative small sample volumes introducing
better approximating a continuous
sample volume uncertainty in laboratory to field scale
media
comparisons
Techniques considered here are
intrusive generating possible
Possible disturbance due to:
Preservation of disturbances due to:
sample quality Fabric changes, stress relief and
Generation of excess pore pressure in
desiccation
saturated soils, unavoidable localised
disturbance,
Ability to vary Possible to simulate e.g. trenching and Not possible to simulate drift in
the properties loading of laboratory samples, environmental conditions and
of the tested parametric sensitivity and long time expensive to simulate long term
material span studies investigations
Time consuming and laborious to
Time (test and Can vary depending on ease of
conduct. Requires both sampling and
interpretation deployment and length of time held
laboratory sample preparation and
duration) static at each sampling point.
measurement.
Can be economic under certain Passive systems are low cost as
conditions, although time delay relative integrated within standard CPT. Active
Cost
to in situ testing may offset any cost systems less widely used, so can be
reduction. costly to mobilise and operate.
Relative infancy in the commercial
sector, though being actively
Equipment Prevalent, well established developed. Future developments must
include robust determination of
accuracy and precision
12
BEST PRACTICE GUIDE
3. Selection of soil thermal properties
3.1 Introduction
This section provides guidance on the selection of design values of thermal properties from different
data sources – ranging from preexisting databases to project specific sets of measurements (Table 3-
1). These data scenarios are illustrative of different project stages (Figure 3-1). As the data scenarios
become more advanced, the design range (or the uncertainty) in thermal conductivity will reduce.
Although each scenario refers to a particular data type, in practice all data types available to the project
should be combined or compared, to take most value from the available measurements.
Figure 3-1: Illustration of data scenarios (Figure 1-4 repeated)

Additional data sets become available during project development
Data
Scenarios
Thermal
conductivity
range
13
BEST PRACTICE GUIDE
Table 3-1: Data scenarios for Thermal Conductivity estimation (Table 1-1 repeated)
Data
scenario / Basis for estimating Thermal
Data type Key data Section
survey Conductivity
stage
No site specific Full thermal database scaled
0 None Section 3.3.1
data for region
Geological
1 Soil type Soil type database subset Section 3.3.2
setting
T-CPT Direct Thermal Conductivity
Sections 2.5
2 In situ testing measurement or measurement or
and 3.4.1
CPT profile Correlated porosity & database
Moisture content, Correlated soil type, porosity & Section 2.3
3a Basic lab testing
specific gravity database and 3.4.2
Thermal Measurement of
Measured soil type, porosity & Section 2.3
3b Conductivity lab Thermal
database and 3.5
test Conductivity
Modelled Effect PSD – to
of trench on estimate change Changed porosity (using Id) &
4 Section 3.6
Thermal in porosity from isomaterial line
Conductivity in situ value
3.2 Thermal property geodatabases

Thermal property geodatabases can be produced based on available data, for instance needle probe
thermal conductivity (TC) measurements, from a range of seabed soils from different projects. These
databases can be used in early phases of the project for pre survey estimation of TC (see Section 3.3).
Detailed information on mineralogy, grain size and other physical parameters (bulk and dry densities,
moisture content, porosity) should be evaluated together with the TC measurement. For instance, poorly
saturated samples (saturation ratio, Sr < 80%) should be excluded from any dataset, provided that
sufficient metadata are available. In case these data are unavailable, the robustness of a geodatabase
could be impacted and the range of correlations that could be explored becomes more limited, also
hampering the evaluation of quality for each data point.
The available data points can be subdivided into five broad soil classifications: (i) quartz sand, (ii)
carbonate sand, (iii) silt, (iv) clay and (v) peat.
3.2.1 Statistical analysis

Statistical analysis should be undertaken on the geodatabase, with standard parameters shown in Table
3-2. An example of presenting the data in scatter plots (combined box, whisker and “beeswarm”) is shown
in Figure 3-2.
14
BEST PRACTICE GUIDE
Outlier detection of the normally distributed data can be calculated using a classic approach to calculate
“outlier fences”, to identify data points that require further investigation (potentially excluded). Where
metadata are available samples should be filtered to remove measurements that do not meet the
selection criteria or sample preparation standards. Measurements without extensive metadata can be
retained, though this might be a source of uncertainties.
Figure 3-2: Combined box, whisker and beeswarm scatter plot - statistical parameters
15
BEST PRACTICE GUIDE
Table 3-2: Normal distribution statistical parameters – geodatabase evaluation
z Definition
MAX TC (W/mK) Maximum value
MIN TC (W/mK) Minimum value
MEDIAN [P50] TC (W/mK) Median, 50th Percentile, middle value
MEAN TC (W/mK) Mean Value
STDEV (W/mK) Standard Deviation
P10 (W/mK) 10th Percentile
P90 (W/mK) 90th Percentile
Upper Quartile (UQ) (W/mK) 75th Percentile
Lower Quartile (LQ) (W/mK) 25th Percentile
Inter Quartile Range (IQR) (W/mK) UQ-LQ
IQR*1.5 (W/mK) Mild whisker magnitude
IQR*3 (W/mK) Extreme whisker magnitude
UQ+IQR*3 (W/mK) Extreme upper whisker threshold – outlier identification
UQ+IQR*1.5 (W/mK) Mild upper whisker threshold – outlier identification
LQ-IQR*1.5 (W/mK) Mild lower whisker threshold – outlier identification
LQ-IQR*3 (W/mK) Extreme lower whisker threshold – outlier identification
N Number of data points
3.2.2 Repeatability-correction of TC database for each soil type

Subsets of any database might include multiple Thermal Conductivity measurements made on the same
soil samples. These repeat tests on sub samples of the same soil provide different measurements of TC,
indicating the repeatability error associated with the test. Based on the review of TC test repeatability, a
coefficient of variance (CoV or v) per soil type can be applied on the raw database values, considering
TC measurements with the same method (e.g. needle probe).
3.2.3 Weighted database according to seafloor soil variation

The database might contain measurements on predominantly specific soil class, e.g. more TC
measurements from sands compared to clays or other soil classes. When considering a single TC value
for initial design phases, the actual distribution of seafloor soils might be different compared to the
database. Hence, a weighting must be applied when combining the different soil types within the
database, rather than considering simply the statistics of the available database.
16
BEST PRACTICE GUIDE
3.2.4 Thermal conductivity versus porosity – trend lines per soil type
The key relationships relevant to the calculation of porosity from standard laboratory measurements are
shown in Table 3-3. These are standard relationships that rely only on the assumption of full saturation
(i.e. they are written here for two phase materials, water and solid, with no gas phase – extensions apply
for partially saturated soils). These relationships provide a route from data of moisture content, 𝑤, and
grain specific gravity, 𝐺𝑠 , to porosity, 𝑛.
Table 3-3: Relevant phase relationships to derive porosity from standard laboratory data
Relationship Formula Notes
𝐺𝑠 can be estimated based on the soil mineralogy or

measured with tests such as the pycnometer (density
bottle).
Grain specific gravity, 𝐺𝑠 𝐺𝑠 = 𝜌𝑠 /𝜌𝑤 There is relatively small variation in 𝐺𝑠 between soil
minerals. Typically values: quartz sand: 2.65, other soil
minerals: 2.6 − 2.8
Grain density, 𝜌𝑠 ; water density, 𝜌𝑤
Voids ratio, 𝑒, from moisture Moisture content: 𝑤 = 𝑊𝑤 /𝑊𝑠 , (weight of water /

content, 𝑤 and grain specific 𝑒 = 𝐺𝑠 𝑤 weight of dry solids
gravity, 𝐺𝑠 Assumption: soil fully saturated
By definition: porosity is pore volume relative to total

Porosity, 𝑛 from voids ratio, 𝑒
𝑛= volume, whereas voids ratio is pore volume relative to
𝑒 1+𝑒
solid volume
The standard geometric relationship between the thermal conductivity of seawater (𝜆𝑓𝑙𝑢𝑖𝑑 : 0.6 𝑊/𝑚𝐾),
the solid phase thermal conductivity (𝜆𝑠𝑜𝑙𝑖𝑑 ) and porosity (n), to estimate the bulk thermal conductivity
range for marine sediments is presented in Equation 1. Typical values for grain thermal conductivities of
marine sediments are presented in Table 3-4.
𝝀𝒃𝒖𝒍𝒌 = 𝝀𝒔𝒐𝒍𝒊𝒅 𝟏−𝒏 𝝀𝒇𝒍𝒖𝒊𝒅 𝒏 Equation 1
17
BEST PRACTICE GUIDE
Table 3-4: Typical grain thermal conductivities of marine sediments

Mineral Phase Grain Thermal Conductivity [W/mK] Source
Quartz 7.7 Horai, 1971
Clauser & Huenges, 1995; Horai,

Calcite 3.5 – 3.6
1971
Horai, 1971; Clauser & Huenges,

K-Feldspar 2.3 – 2.4
1995
Illite 1.8 Midttømme et al., 1998
Montmorillonite* 1.4 Knutsson (1983)
Chlorite 3.1 - 5.25 Clauser & Huenges, 1995
Kaolinite 2.8 Brigaud & Vasseur (1989)
Notes: Measurements taken on compacted bentonite with a montmorillonite composition
When the database measurements correlate well with theoretical relationships between porosity and TC,
the resulting scatter can be used to establish confidence bounds. Hence, new correlations for estimating
TC from porosity and mineralogy can be established, providing initial estimates of thermal conductivity
in the absence of direct thermal measurements. Specifically, Equation 1 can be calibrated to the sub
database of each soil type of the database, by adjusting 𝜆𝑠𝑜𝑙𝑖𝑑 to find a best fit (P50) value, denoted 𝜇𝜆 ,
and evaluating the standard deviation of this parameter, 𝜎𝜆𝑠 . By considering a normal distribution for the
database, the P10 and P90 values of 𝜆𝑠𝑜𝑙𝑖𝑑 can also be calculated.
3.3 Pre survey estimation from databases
3.3.1 No site specific data available (Data Scenario 0)

At early stages of a project, when no site specific information is available (Data Scenario 0 in Figure 3-1
and Table 3-1), the Thermal Conductivity can be estimated by assuming that available thermal
conductivity database(s) of all soil types is relevant.
Within a specific database the scatter due to measurement repeatability error can be removed based on
the repeatability evidence, as described in Section 3.2.2. Moreover, depending on the prevalence of each
soil type at the seafloor across the site (for instance UK waters or any specific region of interest), the
database can be weighted accordingly (Section 3.2.3).
Based on the distribution of sediments, mean and standard deviation values can be calculated by
combining weighted populations. From this, the mean, standard deviation, lower (e.g. P10) and upper
(e.g. P90) percentile values of Thermal Conductivity can be provided (Section 3.2.1).
18
BEST PRACTICE GUIDE
3.3.2 Soil type data available (Data Scenario 1)

In cases where the likely seabed soil type is known (Data Scenario 1 in Figure 3-1 and Table 3-1), without
any other site specific data available, the Thermal Conductivity can be estimated by assuming that the
available thermal conductivity database for that specific soil type is relevant. The corresponding
statistical values can be used, after measurement repeatability error has been removed as described in
Section 3.2.2.
For each soil type the statistics of the repeatability corrected database can be used to provide mean,
standard deviation, lower (e.g. P10) and upper (e.g. P90) percentile values of Thermal Conductivity
(Section 3.2.1).
In summary, for Data Scenario 1, where the seabed soil type is known or estimated, the repeatability
corrected geodatabase statistics provide soil type specific design values.
3.4 Site specific data available excluding thermal properties
3.4.1 In situ CPT data available (Data Scenario 2)

Data Scenario 2 is where site specific in situ CPT testing has been performed (Figure 3-1 and Table 3-1).
Standard CPT data can be used to estimate soil type, and then bulk density and therefore porosity using
correlations based on the CPT parameters (see also Appendix A.2.3). The soil type can be estimated
using conventional Soil Behaviour Type charts (e.g. Robertson charts). From this, a soil mineralogy can
be estimated, potentially using insight from the regional geology, from which the grain specific density
and thermal conductivity can be estimated. Further information is provided in Appendix A.2.3.
Following this step, the Thermal Conductivity can be estimated from thermal conductivity relationships
for two phase materials, as discussed in Section 3.2.4. A similar approach to find thermal conductivity
from CPT data is set out by Vardon & Peuchen (2021).
In summary, for Data Scenario 2, where CPT data is available, a chain of correlations can be applied to
estimate the soil type (including mineralogy) and the porosity. From these values the thermal conductivity
can be estimated using porosity-Thermal Conductivity relationships developed for a specific soil type
(see Section 3.2.4). The repeatability error correction is incorporated in these trend lines (see Section
3.2.2). For other soil mineralogies, different to the database soil types, Equation 1 can be used directly
with the grain mineralogy (with the 𝜆𝑠𝑜𝑙𝑖𝑑 values for various minerals given in Table 3-4).
Where campaigns of direct, in situ, thermal conductivity has been measured, using either passive or
active approaches, then significantly more reliable measurement are available at this stage. As described
in Section 2.5 these measurements are still in their relative infancy in terms of OWF projects and these
techniques still require more research before their accuracy and precision can be fully determined. At
present they should be used in tandem with conventional laboratory based measurements which have
been prepared in a manner that confidently replicates the in situ conditions. The sector should be striving
to ensure that these approaches undergo an extensive multi OWF testing campaign to provide this
confidence.
3.4.2 Lab test data available (Data Scenario 3a)

In Data Scenario 3a, basic soil classification data is available, so the soil type, mineralogy and porosity
are assumed to be known – with a reduced uncertainty compared to the CPT based estimates of Data
Scenario 2 (Section 3.4.1). Porosity can be calculated from measures of bulk density or voids ratio,
following the relationships given in Table 3-3.
19
BEST PRACTICE GUIDE
Following this step, the Thermal Conductivity can be estimated from thermal conductivity relationships
for two phase materials (see also Section 3.4.1), using database trend lines as described in Section 3.2.4.
In summary, for Data Scenario 3a, where basic soil lab test data is available, the thermal conductivity can
be estimated using porosity-Thermal Conductivity relationships developed for a specific soil type
(Section 3.2.4) or grain mineralogy (Table 3-4).
The repeatability error correction is incorporated in the soil type trend lines (see Section 3.2.2).
3.5 Site specific thermal property data available (Data Scenario 3b)
In Data Scenario 3b, project specific thermal conductivity measurements have been obtained. These data
can be used directly to specify design values, taking statistical estimates of Thermal Conductivity for
each soil unit. However, part of the scatter within the measurements is repeatability and the remainder
represents the actual natural variability. The repeatability correction method is also applicable to sets of
data gathered from the same soil unit during a site survey. It is therefore possible to estimate the actual
range of thermal conductivity by correcting the collected data using the same approach (as described in
Section 3.2.2).
It is recommended that from these final measurements, and where appropriate these laboratory based
measurements are combined with in situ directly measured thermal conductivities. The following
statistics should be calculated: Maximum, P90 and P75 values are calculated for each soil unit and used
for sensitivity testing of different cable designs. Reference should be made to the spatial location of the
values in the top 10% of the measurements to identify, from any available ground model, if these values
are actually spatially restricted and/or related to a single soil type. Ultimately, these values should drive
risk based decision making and should not be seen as single values to which a whole windfarm cabling
system is designed. Mitigation strategies such as: changes in cable sizing at different sections of the
route; minor cable rerouting; and shallowing depth of lowering (with or without surface protection) should
be considered.
3.6 Modelling trench effects – allowing for soil disturbance (Data Scenario 4)
Data Scenario 4 represents a situation in which the effect of a trench on the soil thermal conductivity is
to be considered. The database and mixture models (Equation 1) can be used to estimate the effect of
the trenching process, by linking the mechanical effect of the trenching process to the thermal properties.
For sandy soils, the trenching process may loosen the soil, leading to lower thermal conductivity. The
Thermal Conductivity porosity trend lines given by Equation 1 provide a basis to link the loosening
(change in porosity) to the reduction in conductivity. These trend lines represent a single material, with a
given mineralogy (so can be referred to as an iso-material line). As the porosity is altered by the trenching
process, the thermal conductivity will follow the iso-material line. The limits on porosity change from the
trenching process can be estimated through maximum and minimum voids ratio test data, and/or relative
density analysis of CPT data.
For clay soils, the trenching process may produce a more complex situation with blocks of intact clay
surrounded by muddy water or clay slurry. In the long term, the clay will reconsolidate but now in a
normally consolidated condition, at a higher voids ratio (and porosity) compared to the initial condition
– but on the same isomaterial line.
20
BEST PRACTICE GUIDE
These concepts for estimating the thermal conductivity (and permeability) of the trenched material were
explored in WP3 of this project, through numerical modelling (different trenching and burial methods). A
brief summary of key results include:
• The burial process tends to loosen the seabed material, reducing the thermal conductivity,
therefore tending to increase the cable temperature for a given load;
• Ploughing disturbs a larger zone of soil than jetting, so affects the cable temperature more;
• An exception is in stiff clay, which forms a blocky backfill with large voids. If these voids remain
filled with water, then convection is significantly enhanced, leading to significant cooling of the
cable. If the voids fill with sand or reconsolidated clay, then the net effect on cable temperature
could be either an increase or a decrease.
Figure 3-3 illustrates typical results of these effects for sand and soft clay for example installation
scenarios. The predicted changes in cable temperature are specific to the cable and soil properties
modelled in this study, so are only indicative of general trends and should not be relied on for design.
Figure 3-3: Cable temperature change due to burial - induced seabed properties changes
A. Wished-in-place - perfect
Scenario simultaneous lay & burial
B. Pre-cut plough - post-backfilling C. Plough - simultaneous lay & burial
1.5m 1.5m 4.5m
0 < ΔT < |5| OC

1 2 2
2 DoB = DoB =
DoB =
Legend |5| < ΔT < |10| OC 1 1.5m 1.5m
1
1.5m
1
Width =
ΔT > |10| OC 0.3m
1.2m 1.2m 0.8m
Void ratio: 0.49/0.90

Prosity: 0.31/0.47
5.5 OC 7.8 OC 8.9 OC
TC: 3.49/2.30 W/mK
Load: 1400 A
Prosity: 0.36/0.49
4.7 OC 6.5 OC 7.5 OC
TC: 3.11/2.22 W/mK
Load: 1370 A
SAND
Prosity: 0.44/0.55
4.0 OC 5.6 OC 6.4 OC
TC: 2.48/1.91 W/mK
Load: 1300 A
Prosity: 0.50/0.58
3.4 OC 4.8 OC 5.5 OC
TC: 2.15/1.74 W/mK
Load: 1265 A
Prosity: 0.31/0.47
SOFT CLAY 1.6 OC 2.1 OC 2.7 OC
TC: 3.49/2.30 W/mK
Load: 1090 A
21
BEST PRACTICE GUIDE
– Techniques for measuring thermal conductivity for measuring thermal conductivity

BEST PRACTICE GUIDE
Appendix A.1 Laboratory techniques
Appendix A.1.1 Needle probe

The non steady state (transient) needle probe (NSSP) accounts for the vast majority of published thermal
data for geomaterials. It is one of the oldest most established techniques, and the only one with
standards dedicated to geomaterials (ASTM D5334-2014 & IEEE Std-442-2017). Needle probes are
preferred for their short test duration, low levels of sample preparation and the standardised method with
accurate results obtained.
The needle probe test consists of heating a wire in addition to a thermocouple temperature sensor and
is used to solve analytical calculations for the temperature rise at the thermocouple. During
measurement, the probe is inserted and allowed to equilibrate within a sample before a constant power
𝑞 (W/m) is applied for a transient period 𝑡ℎ𝑒𝑎𝑡 (s) inducing a temperature rise (∆θ) in the source.
The present standards outline the factors that influence the results including sample preparation,
experimental configuration, contact resistance, calibration, heat time and power (Section 1.1). Several
geometries and iterations are available for both laboratory and direct field measurements, e.g. instance
TP02/TP08 (Hukseflux) and KD2/KD2-Pro (Decagon).
Figure A.1: Diagram of TP02 needle probe
Notes: 1: Reference temperature sensor pt1000. 2: Heating element (wire) runs across 2/3
needle length. 3: Hot joint thermocouple junction. 4: Cold joint thermocouple junctions. 5:
Power supply.
Typical, experiments demonstrate an initial nonlinear, transient, section (Figure A.2) which reaches a
quasi-steady state after overcoming contact resistance and the thermal properties of the probe
(Vos, 1956). After longer periods the temperature variation with time becomes nonlinear again due to
edge effects, axial heat flow (Vos, 1956) and potentially moisture migration or convective heat transport
(ASTM D5334-14). The influence of convection may be removed by ensuring that the gradient of the
slope is calculated from the linear section only. The cooling section of the curve can also be used in
addition to the heating curve, providing additional results (Low et al., 2015 and 2016).
In deriving the slope, both IEEE Std 442-2017 & ASTM D5334-14 plot the results on a semi-log graph, i.e.
Δ𝜃 against ln 𝑡. The quasi-linear segment is then determined by excluding the first 10-30s for small
diameter probes (ASTM D5334-14) or visually by selecting two appropriate times between which a
straight line can be fitted to the data and its gradient calculated.
BEST PRACTICE GUIDE
The use of visual inspection introduces interpretive variability. Automated analysis of the results through
regression could be beneficial in standardizing interpretation through the selection of the quasi-linear
segment. In (Low et al., 2015 and 2016) the propagation of errors method was used to indicate errors of
12.7% due to uncertainty in gradient evaluation.
To obtain appropriate time ranges for accurate calculations a simulation investigation using finite
element analysis was undertaken based on the TP02 design matching experimental conditions by the
University of Southampton. To define the best calculation periods for different TC ranges, assessment
was performed to highlight the overlaps of low error calculation periods:
• At the lower end of the TC spectrum (0.75 to 1.75W/mK) a 100 seconds measurement window,
with a starting time of ~ 150 seconds is best applied.
• The middle range (1.85 to 2.85W/mK) the low error calculation periods become more
constrained. Periods are closer to the start of the heating cycle, meaning that the transient period
is shorter. A starting time of at least 60 seconds and a calculation window of 130 seconds is
recommended.
• The higher end of the TC spectrum (2.95-3.95W/mK), the range of accurate calculation periods
can have a starting time anywhere between 30–70 seconds and a duration of 10–80 seconds.
Figure A.2: Example of needle probe test results (Low, 2016)
Notes: TC can be determined by using both heating and cooling phases, but the
heating phase is by far the most common approach
Appendix A.1.2 Transient plane source (hot disk)

The transient plane source (Hot Disk) method was developed to determine both TC and diffusivity, which
allows the determination of specific heat capacity of a specimen. Two configurations of the probe are
available, the hot plate/square or the hot disk onto which a thin bifilar spiral, is applied by a deposition
technique or through using thin metal foils enclosed between thin insulating films (BS EN ISO, 22007-2,
2015). This results in the disk acting as both heat source and temperature sensor (Figure A.3).
BEST PRACTICE GUIDE
Figure A.3: The Transient Plane Source Technique (Gustafsson, 1991)
The theory presumes the source is surrounded in an infinite homogeneous medium and the probe is
situated between two identical specimens of the same solid. Once equilibrated, the temperature is
increased by 1 to 2K through the application of a constant current therefore as the temperature changes
so does the resistance. The heat transfer model for the hot disk method does not have a theoretical
analytical solution as it is an absolute technique. Commercially available hot disks come with software
developed to solve partial differential heat transfer equations iteratively.
This method is favoured for its speed and low sample disturbance, accuracy and simultaneous resolution
of TC and thermal diffusivity. Τhere are no standards for applying hot disk to geomaterials, however
(Gustafsson, 1991) demonstrated a 3% uncertainty when applying it to solids. Accurate readings are
predicated on good thermal contact, which for coarser unconsolidated geomaterials becomes difficult.
Appendix A.1.3 Guarded hot plate

The guarded hot plate is a versatile, non intrusive, plane source technique for determining TC with the
direct application of Fourier’s Law. There are two principal configurations although many geometries are
possible: the first sandwiches a hot plate between two samples of the same specimen (Figure A.4a) and
the second shown in (Figure A.4b) sandwiches a single sample of the specimen above a hot and below
a cold plate. Both rely on the axial generation of a steady state temperature gradient, the first up through
the top and down through the bottom samples, the second just axially up through the sample.
The theory is predicated on the development of one directional flow which requires adequate application
of insulating guards to prevent losses at the base and radial flow from the sides. Insufficient insulation
will result in an overestimation for the applied power and consequently the TC. This effect is reduced in
the two specimen configuration.
Guarded hot plate are easy to construct and operate, relying on simple analysis and the generally sample
preparation is minimal. However, the variants of this technique are well established in the determination
of the thermal conductivity of building materials, while for geomaterials, there are uncertainties related
to the equipment, test procedure, sample size and sample preparation. Ensuring construction and
operation follow the specified standards the corresponding accuracies (±2% BSI 12664, 2001),
repeatability (±0.5% BSI 12664, 2001 and ±1% ISO 8302,1991) and reproducibility (±1% BSI 12664, 2001;
ISO 8302,1991) can be achieved. These standards were not written specifically for soils and practicalities
of this approach might be challenging on a project basis.
BEST PRACTICE GUIDE
Figure A.4: The two forms of guarded hot plate method
S Insulation S
e e
c c
Cold surface assembly
o o
n n Cold plate
d Top specimen d Heat
a a Inner
r Primary Primary r
flow Sample heater
guard Hot plate guard
y y
g Bottom specimen g
u u
a Cold surface assembly a
Guarded Hot plate
r r
d Insulation d
(a) (b)
Notes: (a) hot plate sandwiched between two specimens of a sample with steady state temperature
gradient being achieved up through the first specimen and down through the lower specimen (ASTM
C177-19, 2019). (b) is a single sample, heat is generated by the base plate and transmitted up
through the sample to the cold plate generating a steady state temperature gradient
Appendix A.1.4 Thermal cell

The thermal cell is also a non intrusive plane source technique and is theoretically identical to the guarded
hot plate and so can be considered an iteration of this method. It has been specifically developed to
analyse thermal properties of U-100 cores and BS 5930:2015+A1 (2020).
The theory is based on the generation of constant axial (1D) thermal gradient produced by an embedded
heater in the base which dissipates heat from the top (Clarke et al, 2008). Once steady state is achieved,
if power is measurable, thermal conductivity can be derived through the direct application of Fourier’s
Law otherwise, the lumped capacitance method can be used in special circumstances to evaluate power
(Clarke et al, 2008).
This method is suitable for most soils and relies on simple analysis directly suited to U-100 cores.
Although, there are no official standards for the thermal cell or alternatives, the ‘Thermal Pile Design,
Installation & Materials Standards’ recommends the thermal cell as the laboratory method along with
needle probes. There are uncertainties related to the equipment, test procedure, sample size and sample
preparation, noting that the Clarke et al (2008) cell and University of Southampton cell are presently not
recommended over transient methods (Low et al, 2017).
Figure A.5: Schematic of a thermal cell (Low, 2016) based on (Clarke et al., 2008)
Key
Insulation Aluminum
Acrylic Bolts
Soil Thermistor
BEST PRACTICE GUIDE
Appendix A.2 In situ techniques
Appendix A.2.1 Passive and active thermal probes

The most widely known, actively heated, thermal probe system that is available for commercial use is the
Fielax VibroHeat, which can be deployed on a Vibrocorer or integrated into CPT systems above the cone
tip. The device combines passive and active heating phases and is deployed into the seabed penetrating
the soil. During this process (shown in Figure A.6): during penetration the temperature usually rises from
frictional heating but settles on an ambient soil temperature. This allows for relative soil temperature to
be recorded as a function of depth determining the temperature gradient. After allowing the dissipation
of the frictional heat (at least 10 minutes of stabilisation time), a heat pulse with a 20 second duration is
enabled and heats the sensor tube and surrounding soil generating a sharp temperature spike. To
measure the decay of thermal energy, the probe remains in the soil for at least 20 minutes allowing for a
depth dependent result for the TC of the soil. highlights this process.
Figure A.6: Method of deployment of Fielax VibroHeat system (Evensett et al., 2016)
It can be difficult to establish the accuracy and precision for these instruments and often making the
information public is not possible. However, the individual components give temperature measurements
to an accuracy of < 100mK with a precision of ± 1mK.
BEST PRACTICE GUIDE
A second heat probe system used in industry with a similar design is the FERITECH FT580 Heat Flow
Probe. Real time TC, spot temperature and thermal gradient measurements can be produced from this
probe within the seabed. The system is a standalone system with a uniform heat source along the length
of the probe including high density (0.05m resolution) array of thermistors. As mentioned, data can be
collected in real time or collected automatically and stored internally and then extracted when the
instrument is recovered. Due to the nature of this system, it can also be left on the seabed for longer
periods of time in a passive mode recording seasonal changes. Currently there are no published or
publicly available work into assessing the performance of this system.
A further active method for in situ TC measurement is the use of a needle probe that is attached to the
tip of a conventional cone penetrometer (CPT). The device operates and is interpreted in the same way
as a lab based needle probe (see Appendix A.1.1) and is usually of slightly greater dimensions, to be
more robust when used offshore. For example, the probe may be 4-8 mm in diameter and extend by 150-
200 mm beyond the tip of a conventional 35.7 mm diameter CPT. These devices are available for
commercial use, but the passive Thermal-CPT is becoming more commonly used (see Section Appendix
A.2.2).
Based on physical experimental investigation and modelling works (undertaken by Emeana et al, 2016;
Hughes et al, 2015a and Callender et al, 2021) the following should be considered:
• Permeability can play a major role within heat dissipation from HV cables. Especially in high
permeability soils, the main mode of heat transportation is through convection, which if
occurring when undergoing in situ testing would be reflected in higher TC measurements.
• There might be discrepancies in comparison to the laboratory testing results, especially in silty
sand or fine and medium sands with low relative densities (c. < 50%).
• It is considered that the main mode of heat transportation is through convection (for high
permeability soils). Currently there is no theoretical or experimental work into the effects
convection may have with these in situ probes.
• Similarly, to date there have been no systematic investigation into the effects of liquefaction
and soil disturbance and how these events may affect the permeability of the soil which could
impact the measurements (see also remarks on sample preparation for laboratory testing).
Appendix A.2.2 Passive method – thermal CPT

Thermal-CPT (T-CPT), Figure A.7, uses an added temperature sensor on a standard cone penetrometer
and follows the same protocol as a CPT (e.g. ISO, 2021). To measure TC the test is stopped and during
this state of static penetration the temperature decay is recorded and from this the temperature time
series measurements for TC can be calculated (Vardon et al, 2018).
The T-CPT system generally performs best in sandy sediments with sufficient friction to generate a
temperature difference between the ground. TC is calculated using one of the two mathematical
methods presented by Vardon et al (2019) within sufficient time the solution will converge. To generate
robust measurements, the critical terms include (see Figure A.8):
• The temperature change needs to be at least 3°C.
• Measurement time should be between 500 and 1000 seconds.
BEST PRACTICE GUIDE
• At that time after plotting a ln(T-T0) / ln(t) graph the gradient should be -1 (Figure A.8).
• The calibration factor is cone specific and needs to be generated from a 2D axisymmetric model
for the deployed cone, including the precise location of the temperature sensor within the cone.
Prior knowledge of the material being used is needed as the calibration factor changes with the
thermal conductivity of the material being used.
• Accurate calculation of T0, this can be achieved through either ensuring sufficient time for heat
dissipation or using the associated equation.
Figure A.7: Thermal cone penetration test (T-CPT) (Vardon et al., 2018)
temperature sensor
r ~ 20mm
push rod friction sleeve cone tip
It is important to recognise that the T-CPT approach offshore is very new within the offshore industry,
with limited available information to conclude whether that measurements taken in situ would be any
more representative of conditions around an installed cable than those taken in the laboratory. For
instance, there has to be consideration that the pushing of the T-CPT into the seabed may in itself change
the thermal properties, via inducing soil volume changes during both the initial penetration phase and the
actual heat dissipation test.
Figure A.8: Example of T-CPT for recorded TC of 1.97W/mK (Vardon et al., 2019)
BEST PRACTICE GUIDE
Appendix A.2.3 Passive method – correlation with conventional CPT measurements

Standard measurements from CPT data can be used to estimate TC including tip resistance, sleeve
friction and pore water pressure. These parameters are all measured in the CPT tests and can be used
to estimate soil type and mechanical properties. The acquired data correspond to continuous profiles
and input data is typically taken as part of site investigations; hence they represent a potentially cost
effective approach.
The primary measurements during a CPT are (i) tip resistance, qc, (ii) sleeve friction, fs, and (iii) pore water
pressure, u2. These parameters are measured during continuous penetration of the cone at a standard
rate of 20 mm/s. The parameters provide an indirect indicator of the soil mechanical properties. Many
correlations exist to estimate parameters such as soil type, and soil strength and unit weight from CPT
data (Lunne et al. 1997). Using these correlations, CPT data can be used to estimate porosity and soil
type, and thereafter the methods set out in Section 3.2.4 can be applied to estimate TC from porosity and
soil type. The approaches to find porosity and soil type are as follows:
1. Estimate the soil unit weight, using correlations based on the CPT parameters
2. Estimate porosity from unit weight, using an estimated soil grain specific gravity, 𝐺𝑠
3. Estimate the soil type and mineralogy based on the CPT parameters
Based on the calculated porosity and estimated soil type and mineralogy, TC can be estimated following
the method set out in Section 3.2.4.
Vardon & Peuchen (2021) used a similar approach to assess TC based on CPT parameters, and examined
the reliability of this method. Their study provides additional information including simple look-up charts
to estimate thermal properties directly from CPT parameters. Their approach is the same as outlined in
this report, but with different values assumed for the apparent solid fraction thermal conductivity to link
porosity and TC.
The measured CPT parameters are converted to normalised CPT parameters as follows:
(𝑞𝑡 −𝜎𝑣𝑜 ) 𝑃 𝑚
𝑄𝑡𝑛 is the normalised cone resistance, calculated as 𝑄𝑡𝑛 = [ 𝑃𝑎
] / (𝜎′𝑎 ) , where 𝑞𝑡 is the corrected
𝑣𝑜
cone resistance, 𝜎𝑣𝑜 is the in situ vertical total stress at the cone tip, 𝜎′𝑣𝑜 is the in situ vertical effective
stress at the cone tip, 𝑃𝑎 is atmospheric pressure and 𝑚 is a stress exponent related to the soil type.
𝐹𝑅 is the normalised friction ratio, calculated as 𝐹𝑅 = 100 ∙ 𝑓𝑡 /𝑞𝑛 where 𝑓𝑡 is the corrected sleeve friction
and 𝑞𝑛 is the net cone resistance, 𝑞𝑛 = 𝑞𝑡 − 𝜎𝑣𝑜 .
Robertson & Cabal (2010) provided the following correlation that provide soil bulk unit weight, 𝛾, as the
following ratio of the unit weight of water, 𝛾𝑤 :
𝜸
𝜸𝒘
= 𝟎. 𝟐𝟕 𝐥𝐨𝐠 𝑭𝒓 + 𝟎. 𝟑𝟔 𝐥𝐨𝐠 𝑸𝒕𝒏 + 𝟏. 𝟐𝟑𝟔 Equation A-1
Subsequently, an alternative formulation was developed by Lengkeek et al. (2018), using a database that
included soils with organic content, and so captures the lower unit weight of these soils:
𝑄𝑡𝑛,𝑟𝑒𝑓
𝛾 𝛾𝑟𝑒𝑓 log( )
𝑄𝑡𝑛
𝛾𝑤
= 𝛾𝑤
−𝛽 𝐹𝑟,𝑟𝑒𝑓 Equation A-2
log( )
𝐹𝑟
The reference values in this expression that best fit the database of Lengkeek et al. (2018) are:
BEST PRACTICE GUIDE
𝛾𝑟𝑒𝑓 = 19 kN/m3, 𝑄𝑛,𝑟𝑒𝑓 = 150, 𝐹𝑟,𝑟𝑒𝑓 = 30 and 𝛽 = 0.412

The bulk unit weight of saturated soil is linked to the unit weight of the solid and fluid phases via following
relationship:
𝛾
( −1)
𝛾𝑤
𝑛 =1− (𝐺𝑠 −1)
Equation A-3
The grain specific gravity is a parameter that varies only slightly between different mineralogies. If it is
assumed that 𝐺𝑠 = 2.65, then the Robertson & Cabal (2010) approach can be used to estimate porosity
from the CPT parameters as:
𝑛 = −0.164 𝑙𝑜𝑔𝐹𝑟 − 0.218 log 𝑄𝑡𝑛 + 0.857 Equation A-4
In order to estimate TC from porosity, it is necessary to adopt a value of thermal conductivity for the solid
grain fraction, 𝜆𝑠𝑜𝑙𝑖𝑑 . This parameter varies significantly between clays, sands and peat due to their
different mineralogical composition. The soil type and the mineralogical composition may be known
based on the geological setting, lab testing or other data, which supports estimation of 𝜆𝑠𝑜𝑙𝑖𝑑 .
Alternatively, the CPT parameters can be used to estimate the soil type, via methods such as the
approach outlined by Robertson (2016).
Two similar approaches were set out by Vardon & Peuchen (2021) to identify correlations for TC and
volumetric heat capacity for saturated soils. From these relationships theoretical TC correlations can be
developed and similar correlations for volumetric heat capacity. Proposed theoretical correlations from
standard CPT measurements are shown in Figure A.9. These have been tested at 7 locations from both
in situ T-CPT data and needle probe measurements on the specific materials which is recognised as a
very small number of samples. There is a level of uncertainty surrounding them, as CPT measurements
are influenced by multiple soil properties. There is no strong bias correlating the results but still a high
scatter suggesting the initial tests are producing accurate but not precise measurements.
Figure A.9: Correlations for TC and CPT based on Vardon and Peuchen (2021)
BEST PRACTICE GUIDE
- References
ASTM International: ASTM C177-19 (2019). Standard Test Method for Steady-State Heat Flux
Measurements and Thermal Transmission Properties by Means of the Guarded-Hot-Plate
Apparatus’, Annual Book of ASTM Standards, (C), pp. 1–23.
ASTM International C1044-12. (2012). Standard Practice for Using a Guarded-Hot-Plate Apparatus or
Thin-Heater Apparatus in the Single-Sided Mode. ASTM International, West Conshohocken, PA.
ASTM International: D5334–14 (2014). Standard test method for determination of thermal conductivity
of soil and soft rock by thermal needle probe procedure. ASTM International, West Conshohocken.
Brigaud, F. and Vasseur, G. (1989). ‘Mineralogy, porosity and fluid control on thermal conductivity of
sedimentary rocks’, Geophysical Journal International, 98(3), pp. 525–542.
British Standards Institution (2001a). BS EN 12664:2001 Thermal performance of building materials and
products – Determination of thermal resistance by means of guarded hot plate and heat flow meter
methods – Dry and moist products of medium and low thermal resistance. BSI, London.
British Standards Institution (2015). BS EN ISO 22007–2:2015 Plastics – Determination of thermal
conductivity and thermal diffusivity – Part 2: Transient plane heat source (hot disc) method. BSI,
London.
British Standards Institution: BS 5930:2015+A1 (2020). Code of Practice for Site Investigation. BSI,
London.
Callender, G. et al. (2021) ‘Low computational cost model for convective heat transfer from submarine
cables’, IEEE Transactions on Power Delivery, 36(2), pp. 760–768. doi:
10.1109/TPWRD.2020.2991783.
Clarke, B. G., Agab, A. and Nicholson, D. (2008). ‘Model specification to determine thermal conductivity
of soils’, Proceedings of the Institution of Civil Engineers: Geotechnical Engineering, 161(3), pp. 161–
168.
Clauser, C., & Huenges, E. (1995). Thermal conductivity of rocks and minerals. Rock physics and phase
relations: a handbook of physical constants, 3, 105-126.
Decagon Devices, Inc. (2011). KD2 Pro Thermal Properties Analyzer: Operator’s Manual Version 10
Emeana, C.J., Hughes, T.J., Dix, J.K., Gernon, T.M., Henstock, T.J., Thompson, C.E.L. and Pilgrim, J.A.,
(2016). The thermal regime around buried submarine high-voltage cables. Geophysical Journal
International, 206(2), pp.1051-1064.
Evenset, G., Klaer, L.E., Kingman, K., vd Thillart, R., Daniela, D., de Vries, G.T. and Dillon, M., (2016). Thermal
characterization of seabed along the NordLink cable route–results and comparison of measurement
methods. In CIGRE 2016 Conference Proceedings. B1-313.
GCube (2021). Uncharted Waters – Navigating emerging risks and rising claims in global offshore wind.
Q4 2021 Report. GCube Insurance.
Gustafsson, S. E. (1991). ‘Transient plane source techniques for thermal conductivity and thermal
diffusivity measurements of solid materials’, Review of Scientific Instruments, 62(3), pp. 797–804.
BEST PRACTICE GUIDE
Horai, K. I. (1971). Thermal conductivity of rock‐forming minerals. Journal of geophysical research, 76(5),
1278-1308.
Hughes T.J., Henstock T.J., Pilgrim J.A., Dix J.K., Gernon T.M. and Thompson C.E.L. (2015a). Effect of
sediment properties on the thermal performance of submarine HV cables. IEEE Transactions on
Power Delivery 30(6): 2443-2450. doi: 10.1109/TPWRD.2015.2398351.
Hughes, T.J., Henstock, T.J., Pilgrim, J.A., Dix, J.K., Gernon, T.M. and Thompson, C.E.L. (2015b). Thermal
Ratings of Submarine HV Cables Informed by Environmental Considerations. Jicable, Palais de
Congres, Versailles, France (21st-25th Jun, 2015).
Hukseflux Thermal Sensors (2011). ‘TP02 Non-Steady-State Probe for Thermal Conductivity URL:
Measurement’. Measurement – manual v0908, Hukseflux Thermal Sensors, Delft.
http://www.hukseflux.com/products/thermalConductivity/tp02.html
IEC 60287-1-1: 2006+A1 (2014). Electric Cables – Calculation of the Current Rating, IEC, Part 1-1 Current
rating equations (100% Load factor) and calculation of losses - General.
IEC 60853-2 (2008).Calculation of the cyclic and emergency current rating of cables – Part 2 Cyclic rating
of cables greater than 18/30 (36) kV and emergency rating for cables of all voltages.
IEEE 442-2017 (2018). IEEE Guide for Thermal Resistivity Measurements of Soils and Backfill Materials.
International Organization for Standardization, ISO 22475-1:2021 (2021). Geotechnical investigation and
testing - Sampling methods and groundwater measurements - Part 1: Technical principles for
execution’, 3, ISO, Geneva.
International Organization for Standardization (1991). ISO 8302:1991 Thermal insulation – Determination
of steady-state thermal resistance and related properties – Guarded hot plate apparatus, ISO,
Geneva.
Knutsson, S. (1983). On the thermal conductivity and thermal diffusivity of highly compacted bentonite.
SKB Report, Swedish Nuclear Fuel and Waste Management Co., SKB 83 – 72, October 1983.
Lengkeek H.J., de Greef J. & Joosten S. (2018). CPT based unit weight estimation extended to soft
organic soils and peat. Cone Penetration Testing 2018 - Proceedings of the 4th International
Symposium on Cone Penetration Testing, CPT 2018, , pp. 389-394.
Low, J. (2016). Thermal conductivity of soils for energy foundation applications (Doctoral dissertation,
University of Southampton).
Low, J. E. et al. (2015). ‘A comparison of laboratory and in situ methods to determine soil thermal
conductivity for energy foundations and other ground heat exchanger applications’, Acta
Geotechnica, 10(2), pp. 209–218.
Low, J. E., Loveridge, F. A. and Powrie, W. (2017). ‘Error analysis of the thermal cell for soil thermal
conductivity measurement’, Proceedings of the Institution of Civil Engineers: Geotechnical
Engineering, 170(3), pp. 191–200.
Lunne, T., Robertson, P.K. and Powell, J.J.M. (1997). Cone Penetration Testing in Geotechnical Practice.
Blackie Academic & Professional, London, 312 p.
Midttømme, K., Roaldset, E., & Aagaard, P. (1998). Thermal conductivity of selected claystones and
mudstones from England. Clay Minerals, 33(1), 131-145.
BEST PRACTICE GUIDE
Pilgrim, J.A., Lewin, P.L., Larsen, S.T., Waite, F. and Payne, D. (2012). Rating of cables in unfilled surface
troughs. IEEE transactions on power delivery, 27(2), pp.993-1001.
Robertson P.K. (2016). Cone penetration test (CPT)-based soil behaviour type (SBT) classification
system — an update. Canadian Geotechnical J. 53: 1910–1927 dx.doi.org/10.1139/cgj-2016-0044.
Vardon, P.J., Baltoukas, D. and Peuchen, J., (2018). June. Thermal Cone Penetration Test (T-CPT). In 4th
International Symposium on Cone Penetration testing. CRC Press. 649-655.
Vardon, P.J., Baltoukas, D. and Peuchen, J., (2019). Interpreting and validating the thermal cone
penetration test (T-CPT). Géotechnique, 69(7), pp.580-592.
Vardon, P.J. & Peuchen, J. (2021). CPT correlations for thermal properties of soils. Acta Geotech. 16,
635–646.
Vos, B. H. (1956). ‘Measurements of thermal conductivity by a non-steady-state method’, Applied
Scientific Research, Section A, 5(6), pp. 425–438.
BEST PRACTICE GUIDE
The Carbon Trust is a company limited by guarantee and registered in England and Wales under Company number 4190230
with its Registered Office at: Level 5, Arbor, 255 Blackfriars Road, London, England, SE1 9AX. © The Carbon Trust 2024. All
rights reserved. Published in the UK: 2024

2024-Soil Thermal

Uploaded by

Copyright:

Available Formats

2024-Soil Thermal

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

2024-Soil Thermal

Uploaded by

Copyright:

Available Formats

Soil Thermal

Properties (STP) for

© The Carbon Trust 2024

2. Thermal soil property collation and measurement 7

3. Selection of soil thermal properties 13

Figure 1-3: Continuous static cable rating for a generic AC cable

1.3 Characterising the seabed thermal environment for cable design

Figure 1-4: Illustration of data scenarios

Table 1-1: Data scenarios for Thermal Conductivity estimation

No site specific Full thermal database

Direct Thermal Conductivity

Moisture content, Correlated soil type,

Modelled Effect of PSD – to estimate

2. Thermal soil property collation and measurement

2.1 Thermal properties within the site investigation process

2.2 Methods for measuring thermal conductivity

2.3 Laboratory testing techniques

• Short test duration • Ensuring good ASTM D5344 (2014):

2.4 Sample preparation

2.5 In situ measurement techniques

Table 2-2: Laboratory based versus in situ geotechnical tests

Laboratory based measurement In situ measurement

3. Selection of soil thermal properties

Figure 3-1: Illustration of data scenarios (Figure 1-4 repeated)

3.2 Thermal property geodatabases

3.2.1 Statistical analysis

Table 3-2: Normal distribution statistical parameters – geodatabase evaluation

3.2.2 Repeatability-correction of TC database for each soil type

3.2.3 Weighted database according to seafloor soil variation

𝐺𝑠 can be estimated based on the soil mineralogy or

Voids ratio, 𝑒, from moisture Moisture content: 𝑤 = 𝑊𝑤 /𝑊𝑠 , (weight of water /

By definition: porosity is pore volume relative to total

𝝀𝒃𝒖𝒍𝒌 = 𝝀𝒔𝒐𝒍𝒊𝒅 𝟏−𝒏 𝝀𝒇𝒍𝒖𝒊𝒅 𝒏 Equation 1

Table 3-4: Typical grain thermal conductivities of marine sediments

Quartz 7.7 Horai, 1971

Clauser & Huenges, 1995; Horai,

Horai, 1971; Clauser & Huenges,

Illite 1.8 Midttømme et al., 1998

Montmorillonite* 1.4 Knutsson (1983)

Chlorite 3.1 - 5.25 Clauser & Huenges, 1995

Kaolinite 2.8 Brigaud & Vasseur (1989)

Notes: Measurements taken on compacted bentonite with a montmorillonite composition

3.3 Pre survey estimation from databases

3.3.1 No site specific data available (Data Scenario 0)

3.3.2 Soil type data available (Data Scenario 1)

3.4 Site specific data available excluding thermal properties

3.4.1 In situ CPT data available (Data Scenario 2)

3.4.2 Lab test data available (Data Scenario 3a)

1.5m 1.5m 4.5m

0 < ΔT < |5| OC

Void ratio: 0.49/0.90

– Techniques for measuring thermal conductivity for measuring thermal conductivity

Appendix A.1 Laboratory techniques

Appendix A.1.1 Needle probe

Figure A.1: Diagram of TP02 needle probe

Figure A.2: Example of needle probe test results (Low, 2016)

Appendix A.1.2 Transient plane source (hot disk)

Figure A.3: The Transient Plane Source Technique (Gustafsson, 1991)

Appendix A.1.3 Guarded hot plate