Guidelines For The Quantitative Description of Discontinuities For Use in Discrete Fracture Network Modelling

Innovations in Applied and Theoretical Rock Mechanics Editors – Hassani, Hadjigeorgiou, Archibald
©2015 by the Canadian Institute of Mining, Metallurgy & Petroleum and ISRM, ISBN 978-1-926872-25-4
13th International Congress of Rock Mechanics ISRM13
10-13 May 2015, Montreal, Canada
Downloaded from http://onepetro.org/isrmcongress/proceedings-pdf/CONGRESS13/All-CONGRESS13/ISRM-13CONGRESS-2015-246/1473531/isrm-13congress-2015-246.pdf by University of British Columbia user on 09 December 2021

GUIDELINES FOR THE QUANTITATIVE DESCRIPTION OF DISCONTINUITIES FOR USE
IN DISCRETE FRACTURE NETWORK MODELLING
*D. Elmo
University of British Columbia
6350 Stores Road
NBK Institute of Mining Engineering
Vancouver, BC V6T1Z4
(*Corresponding author: [email protected])
D. Stead
Simon Fraser University
Vancouver, Canada
S. Rogers
Golder Associates ltd.
Vancouver, Canada

D. Elmo, D. Stead, S. Rogers 1

ABSTRACT
Discrete Fracture Network (DFN) modelling has increasingly being used in many geotechnical
and mining engineering problems, the authors believe there is an ever greater demand to provide updated
guidelines that address the collection of discontinuity data in the specific context of DFN models. In
general terms, the generation of a DFN model requires collecting information on i) fracture orientation, ii)
fracture intensity, iii) fracture length and iv) fracture terminations. The International Society of Rock
Mechanics (ISRM) suggested methods for the quantitative description of discontinuities in rock masses
directly include most of these parameters. However, there are important differences engineers should be
aware. To what extent discontinuities can be sampled and which limitations are inherently introduced in
the analysis by the sampling methods being adopted represent important aspects that should drive the
collection of discontinuity data for DFN analysis; this will require the introduction of a more appropriate
set of guidelines bridging the gap between the current, very practical, ISRM methods and the data
requirements imposed by the use of new DFN technologies.
KEYWORDS
Characterisation, discontinuities, discrete fracture network, mapping of discontinues
INTRODUCTION
The last decade has seen a major increase in use of the Discrete Fracture Network (DFN)
approach, both as a stand-alone tool or integrated within more complex geomechanical simulations (Rogers
et al., 2014 and Elmo et al., 2014). In particular, the DFN approach offers the opportunity to maximise the
use of fracture data collected from mapping of rock exposures and to construct synthetic rock mass (SRM)
models. However, the potential of DFN based modelling is limited if there is insufficient care in collecting
the necessary structural data at the required engineering scale. DFN models are also subject to the process
of data calibration and validation (Hadjigeorgiou, 2012). With DFN modelling becoming an almost
integral part of many geotechnical and mining engineering problems, the authors believe there is an ever
greater demand to review the process of collecting fracture data in the specific context of DFN modelling.
The value of the DFN model depends directly on the quality and quantity of available field data.
For example, fracture length (persistence) is an important parameter in DFN modelling; however, this
parameter is either seldom available at the pre-feasibility stage due to a lack of exposures (man-made or
natural), or engineers have access to limited length data collected along exploratory drifts. Characterisation
of rock exposures should also take into account limits of window/scanline mapping techniques with respect
to both trace length bias and the effect of cut-off assumptions inherent in the mapping methodology.
The International Society of Rock Mechanics (ISRM) suggested methods for the quantitative
description of discontinuities in rock masses (ISRM, 1981) include most, if not all, of the main parameters
required to generate a DFN model. However, there are important differences engineers should be aware of.
For instance, the ISRM methods use fracture spacing, whereas the fracture intensity parameter in the DFN
method can refer to linear, areal and volumetric intensity. The fracture radius (DFN input) does not
necessarily correspond to fracture length or the trace length mapped on an exposed surface. The definition
of fracture radius also requires geotechnical engineers to be familiar with truncation and censoring biases
(i.e. values below a certain fracture length are omitted - truncation - or relatively larger values cannot be
measured because of the limited extent of the rock exposure - censoring). Truncation bias also plays a
major role in defining the correct fracture intensity for DFN analysis. Fracture terminations are seldom
collected by geotechnical engineers in the field, but provide a hierarchal structure to relate discontinuity
sets. To what extent discontinuities can be sampled and which limitations are inherently introduced in the

2 D. Elmo, D. Stead, S. Rogers

analysis by the sampling methods being adopted represents important aspects that should drive the
collection of discontinuity data for DFN analysis.
In order to set a common reference with respect to terminology and avoid ambiguity, the terms
“discontinuity” and “fracture” are herein used as synonyms to define structural features independently of
their genesis. Likewise, the term “fracture length” is used in lieu of the term “fracture persistence” to
define the extent of a fracture within a plane.
PRINCIPLES OF DFN MODELLING
The basis of the DFN approach is the characterisation of each discontinuity set within a structural
domain using statistical distributions to describe variables such as spatial location, orientation, intensity
and size. A direct examination of rock mass structures in three-dimensions (3D) is not possible, and the
DFN approach maximises the use of discontinuity data from mapping of exposed surfaces and boreholes or
any other source of spatial information. Discontinuity data sampled from exposures in variably oriented
outcrops (2D) and boreholes (1D) can be used to generate a 3D stochastic discontinuity model that shares
the statistics of the samples and allows for the incorporation of specific (deterministic) discontinuities, such
as faults or very persistent joints.
Relatively undisturbed rock core samples can be obtained by high quality drill coring; however,
specific sampling and analytical techniques are needed to measure the orientation of the sampled
discontinuities within the rock mass. Observations of exposed rock faces, at or near the project site, have
the advantage of allowing direct measurements of discontinuity orientation, spacing, length and the
identification of discontinuity sets. Other large-scale geometrical and structural features can be readily
observed. Different techniques can be employed, including: scanline survey (linear mapping), rectangular
window mapping and circular window mapping. Increasingly, digital photogrammetry and laser scanning
techniques (LiDAR) provide an alternative technique for fracture characterisation. An increasing trend
toward auto-detection of discontinuities may in turn bring increased uncertainty in DFNs if care is not
taken in the data characterisation process.
The typical process involved in the generation of a DFN model requires the definition of a number
of fracture properties, see Table 1. The primary properties are required for all modelling as these represent
the geometry of the fracture network. However the secondary properties may also be defined depending on
whether the application of the DFN model is hydraulic or geomechanical. In order to build a volumetrically
simple DFN model, only the primary fracture properties are required. Validation of the DFN model is
achieved by comparing the orientation, intensity and pattern of the simulated fracture traces with those
measured in the field. The stochastic nature of the process is such that there are an infinite number of
possible realisations of the 3D fracture system based on the mapped data. Indeed, the mapping process is
itself random by the nature of how fractures are presented in available windows. With the exception of
fully explicit modelling of an individual fracture or simplified fracture sets, the stochastic approach
provides the best option for creating realistic geometric models of fracturing.
Table 1 – Primary and secondary properties for defining a DFN model

Primary Secondary
Orientation distribution Aperture distribution
Fracture length distribution Fracture shear strength properties
Fracture intensity distribution Fracture stiffness properties
Termination percentage Transmissivity distribution
Spatial variation Storativity distribution
Termination percentage


The starting point for the generation of a DFN model is the definition of the spatial model that
governs the way fractures are generated within a given 3D volume. The majority of the spatial models
involve similar considerations for specific fracture characteristics, such as shape (generally polygons), size
and termination at intersections. A comprehensive review of spatial models is outside the scope of this
paper. Literature sources for spatial DFN models include, amongst the others, Dershowitz and Einstein
(1988) and Staub et al. (2002). The main differences lie in the specific distribution laws used to simulate
fracture orientation and fracture location. Spatial models can be grouped in stationary and non-stationary
models based upon the relationship between different fracture sets (non-stationary models incorporate the
option of generating spatial and temporal relations between different fracture sets). The choice of a specific
DFN model is typically based on assumptions made from field data and geological observations. The
Enhanced Baecher model provides an appropriate solution for different applications and is well suited for
relatively small scale DFN models; however, other models that allows for consideration of geostatistical
data in DFNs may be appropriate for more complex, large scale DFNs (e.g. geocellular models in which
fractures are generated to a volumetric grid, based upon properties of the grid).
SAMPLING VARIABILITY AND UNCERTAINTY
The process of characterisation and quantification of rock engineering parameters is strongly

linked to the concepts of variability and uncertainty (Harrison, 2012). Uzielli (2008) defined variability as
the observable manifestation of heterogeneity of physical parameters and/or processes. The same author
defined uncertainty as relating to the engineer’s state of knowledge, reflecting the decision to recognise and
describe the observed variability in a qualitative or quantitative manner. Therefore, the two terms are not
synonyms and should be applied to specific characteristics (Harrison, 2012). With respect to sampling of
natural discontinuities, the randomness of natural geological processes is responsible for the manifested
variability of the parameters used to describe natural discontinuities. Engineers generally assume that such
variability can be described using stochastic (statistical) models; however, this may not be true for all
conditions.
Mapping of natural discontinuities can be carried out according to either a subjective (biased) or
objective sampling approach (ISRM, 1981), independently of whether the discontinuity parameters are
collected using direct physical mapping or indirect data collection by remote sensing techniques (e.g.
digital photogrammetry and laser scanning techniques). Subjective sampling considers only those
discontinuities that are deemed important by the engineer carrying out the sampling. Conversely, objective
sampling considers all natural discontinuities intersecting a given line or contained within a given area.
Therefore, only the latter is a random process, and accordingly the data being collected could be
characterised using statistical analysis, Figure 1. Sampling of natural discontinuities should include
procedures to capture total uncertainty, defined as the combination of aleatory variability, i.e. variability
inherent in random processes, as well as epistemic uncertainty, i.e. lack of knowledge (Elmouttie and
Poropat, 2011).
For subjective sampling any variation in the mapped data would be considered the result of
limited knowledge, and accordingly the data should be characterised using epistemic models (i.e. not
stochastic). Objective sampling of a given rock exposure by scanline or window mapping techniques
allows to move from an epistemic to a stochastic (aleatory) model, Figure 2. Note that increasing the
population of sampled discontinuities does not reduce variability, since variability is irreducible, but
increasing the degree of knowledge improves our understanding of variability (Harrison, 2012). The basis
of DFN modelling is using statistical distributions to describe discontinuity sets, therefore DFN models are
by definition stochastic, and their applicability appears to be confined within the limits of aleatory
variability as “Objective DFN models”. However, DFN models could in principles still be generated using
limited, subjective, data by assuming statistical distributions for the sampled parameters. Since these
“Subjective DFN models” are built upon limited knowledge, they have an inherently larger degree of
uncertainty, which can only be minimised by improving the data collection procedure (quantity and quality
of data being collected).


It is important to realise that Subjective DFN Models may, in the worst case, misrepresent the
behaviour of the rock mass, with great implications for stability analysis, fragmentation analysis, etc.
However, as also recognised by Elmouttie and Poropat (2011), engineers should be aware that adopting
deterministic or equivalent continuum methods in lieu of Subjective DFN Models (on the basis of the larger
uncertainty potentially associated with them) may provide a false sense of security and make engineers
overlook problems in which fracture properties need to be modelled discretely.
Figure 1 – Sampling approach to collect properties of natural discontinuities
Figure 2 – Uncertainty and variability concepts (modified from Harrison, 2012) integrated with sampling
of natural discontinuities and DFN modelling


ISRM QUANTITATIVE DESCRIPTION OF DISCONTINUITIES AND BASIC DFN INPUTS
The ISRM suggested methods for the quantitative description of discontinuities in rock masses
(ISRM, 1981) lists the ten parameters generally recorded when sampling discontinuities in-situ. Amongst
those, the followings are required to develop a basic DFN model:
• Fracture Orientation: for example measured as dip/dip direction to describe the attitude
of the discontinuity in space.
• Fracture Spacing: defined as the perpendicular distance between adjacent discontinuities.
For a scanline survey it represents the progressive distance along the scanline, whilst in
the case of window mapping it is usually given as the perpendicular distance between
discontinuities inferred to belong to the same set.
• Fracture Length (persistence): the trace length of a discontinuity as observed in the
mapped exposure.
No direct reference is made in ISRM (1981) to fracture terminations as a required parameter to

describe discontinuities and rock masses. However, practitioners typically include fracture terminations in
their mapping sheets, by means of which the nature of the termination of each discontinuity trace is
measured (e.g. discontinuity trace terminating in rock, against another discontinuity, or extending outside
the mapped region). Fracture terminations are important parameters governing fracture connectivity, which
is linked to the potential for block formation. Note that ISRM (1981) does include the number of joint sets
and block size among the required parameters to be sampled.
Fracture Orientation
In general terms, no difference exists in the way fracture orientation data should be collected for
DFN purposes compared to the ISRM (1981) suggested methods. Fracture orientation can be sampled
using boreholes (from direct measurements on rock cores or from the acoustic televiewer, ATV). DFN
models can be generated separately for each fracture set and then combined to obtain the overall
representation of the fracture network. The application of separate statistical procedures to define fracture
sets is known as a disaggregate approach (each fracture sets is also defined separately with respect to the
statistical analyses for fracture intensity and fracture radius).
Distributions such as Fisher, Bingham, bivariate Fisher and bivariate Bingham can be used to
represent fracture orientation. Alternatively, field data that do not conform to straight forward statistical
methods (i.e. characterised by a highly dispersed scatter), can be analysed using a bootstrap approach,
whereby a statistical method based upon multiple random sampling with replacement from an original
sample is used to create a pseudo-replicate sample of fracture orientations (Rogers et al., 2014). Since
orientation data are invariably biased, orientation distribution should be corrected to account for fractures
intersecting the borehole at a low angle to be misrepresented. However, when validating the DFN model,
the simulated orientation data would have to be compared to the uncorrected data, since the sampling
procedure would automatically introduce equivalent sampling bias.
Fracture Spacing (Intensity)
The ISRM parameter “fracture spacing” is not used as such in DFN modelling, which relies on
the definition of fracture intensity parameters. However, there are a wide range of possible measures, often
with ambiguous definitions, that can be used to define fracture intensity. Dershowitz and Herda (1992)
introduced a unified system of fracture intensity measures that provide engineers and DFN users with an
easy framework to move between differing scales and dimensions. Fracture intensity is referred to as Pij
intensity, where the subscript i refers to the dimensions of sample, and subscript j refers to the dimensions
of measurement. Accordingly, the volumetric fracture intensity (P32) is defined as the ratio of total fracture
area to unit volume (dimensions of m2/m3). P32 is an intrinsic rock mass property and whilst it cannot be
directly measured, it can be inferred from either 1D or 2D data using a simulated sampling methodology on


the basis of a linear correlation with P21 (m/m2), which is the total trace length of fractures per unit area, or
with P10 (m-1), which represents the total number of fractures along a scanline or borehole (i.e. P10 is a
measure of fracture frequency).
P32 = C32P21 (1)
P32 = C31P10 (2)
Because both P10 and P21 data are heavily influenced by the relative orientation of the
predominant fracture sets with respect to the orientation of the boreholes/sampling planes, the constants of
proportionality C32 and C31 also depend on the relative orientation of the fractures to the orientation of the
sampling panel or scanline/borehole. Furthermore, by definition, the volumetric fracture intensity P32 is
related to fracture radius, thus truncation biases would play a major role in defining the correct fracture
intensity for DFN analysis.
The primary input for DFN modelling of fracture intensity is borehole derived fracture frequency
(P10) data. By using cumulative fracture intensity (CFI) plots, each borehole can be interpreted to identify
zones of the rock mass where P10 remains constant over given intervals lengths (interval lengths would
depend on the modelling resolution, small scale or large scale DFN models). CFI plots display depth on the
vertical axis and cumulative fracture frequency on the x-axis. Where the gradient of the CFI curve is
relatively constant, the fracture frequency (P10) over that interval is constant and can be determined.
Interpretation of CFI plots from a large number of boreholes results in the creation of a data set of specified
P10 values and interval lengths that provide the basis for any 3D spatial modelling and extrapolation of
fracture intensity (Rogers et al., 2014).
Once the data set of specified P10 values and interval lengths is available, DFN models can be
generated by conditioning the model to a direct replication of the number of fractures intersected along a
scanline/borehole. This process is known as P10 conditioning. To account for directional bias, P10
intensity properties can be converted to a non directional P32 potential property using the method proposed
by Wang (2006) using stereological relationships between fracture orientation and fracture intensity. The
P32 potential is independent of scale and orientation, and can therefore be used to directly extrapolate
intensities throughout the rock mass volume. This method is based on several key assumptions: i)
assumption of 1D line sampling (i.e. a zero-radius borehole), which implies that every fracture (for a given
borehole) crosses the entire diameter of the core; and ii) more importantly, the method assumes that the
fracture population in a single set follows a univariate Fisher hemispherical probability distribution (the
method is not suitable for use with other types of probability distributions.
Fracture Length
The definition of fracture length in the context of DFN modelling requires acknowledging the
critical difference between fracture length (mapped on rock exposures) and fracture radius, Figure 3.
Fracture length is an explicit measure of the trace that a fracture or fault makes within a geological surface
or mining exposure. Fracture radius is the radius of a circle of equivalent area to a polygonal fracture.
Fracture length doe not directly correspond to fracture radius since trace lengths observed on tunnel walls
or bench faces are not actually diameters, but may be cords to larger discs (assuming fractures are circular).
Measurements of fracture length can be obtained by mapping 2D rock exposures, using conventional
(scanline or window mapping) or remote sensing techniques (e.g. photogrammetry and/or LiDAR). Note
that generally remote sensing techniques process fracture length as the diameter of a disc inscribing the
mapped feature. Therefore, it is very important not to confuse the notion of fracture radius used in the DFN
model with the apparent radius of the disc inscribing the fracture plane in the processed image or point
cloud.
Once the distribution of fracture length is known, the distribution of fracture radius can be
assessed by analytical methods (Mauldon, 1998; Zhang and Einstein, 2000). The process is based on the


basic assumptions that: i) fractures are considered to be planar; and ii) fractures can be represented by thin
polygons with n sides (i.e. fracture are assumed to be equidimensional). However, there exists the
possibility in some DFN modelling packages to account for non equidimensional fractures. The analytical
method is well suited for deriving the distribution of fracture radius using the distribution of fracture
lengths as the initial input. The method is based on the analysis of data mapped within circular windows,
which have the advantage over scanline and rectangular windows of automatically eliminating orientation
bias. Note that this method may require collecting information about specific types of fracture terminations
(see also section below). By mapping fractures in-situ using techniques that allows for digitizing traces,
engineers could then use automatic algorithms to process the sampled information, see Figure 4, with
termination types accounted for automatically.
Figure 3 – Distribution of mapped fracture length (data source) versus distribution of fracture radius
(required DFN input)
Figure 4 – Recommended method to collect fracture length (either surface or underground). The method
can be implemented within DFN based codes to automatically account for censoring and truncation biases


Data Collection, Variability and Perceived Uncertainty
With respect to data collection, it is important to consider the limitations that may arise when
sampling fracture length (persistence) data using the modal intervals suggested in ISRM (1981). As shown
in Figure 5, an otherwise an aleatory data set may be perceived as an epistemic data set because of the
subjectivity of the data characterisation process. When collecting fracture length data for DFN analysis it is
recommended not to lump data into pre-defined lengths intervals, rather to measure each length separately
and then process the full data using length intervals sets that allow for a better characterisation of the
underlying distribution. The choice of the distribution often depends on the modelling scale; for instance,
small scale fractures and large scale faults could be related according to power law distributions, whereas
other forms of distributions (e.g. log normal) could provide a better form of length characterisation for
small scale DFN models (e.g. tunnel scale or bench scale).
Reconciling Fracture Size and Intensity
The main issue to consider when mapping fracture length for DFN analysis is that any fracture
intensity measure would have an associated minimum fracture size associated with it. Accordingly, the
probability density function for fracture size is bounded by a minimum (estimated at the borehole scale)
and a maximum (estimated at the mapping scale); the mapped intensity reflects the extent of the size
distribution available. For example, as shown in Figure 6, assuming a P32 of 1m-1 represents the fracture
intensity for a minimum length of 0.1m (no maximum), the P32 for the range [2m, 30m] would be 0.23m-1.
A truncation or size cut-off is generally assumed when mapping fractures using either scanline or window
sampling methods, while the maximum cut-off may be imposed by the geometry of the problem (e.g. in
underground mining the maximum length is associated with the height of the excavation). Practitioners
should be aware that the fracture frequency (intensity) measured in the field refers specifically to the length
intervals being mapped, and should not be used to generate an unbounded (length wise) DFN model. For
the DFN model to include lengths outside the mapped intervals then fracture intensity should be corrected
accounting for the type of length distribution and new cut-off values.
Figure 5 – Example of the limitations that may derive when sampling fracture length (persistence) data
using subjective length intervals, as for example those suggested in ISRM (1981), left chart. The figure
clearly shows that an otherwise an aleatory data set may be perceived as an epistemic data set based on the
subjective way in which length data are lumped together.


Figure 6 – Effects of adjusting the minimum (and maximum) size (radius) truncations on the total
probability density function of fracture intensity
Fracture Terminations
Fracture termination is a parameter that is generally overlooked when collecting data for rock
engineering problems, and not explicitly considered in ISRM (1981). Fracture termination (in intact rock,
at another joint, in rock across another joint) relates to fracture connectivity, which is an important
parameter in the context of both permeability and stability (through failure of intact rock bridges). Fracture
termination could provide useful insights on the structural character of the rock mass, as well as it could be
used as a calibration tool by comparing mapped terminations on exposed surfaces with those generated in
equivalent surfaces within the DFN models. One type of termination (Figure 7) which is very important is
the so called T-type, which relates to the order in which fracture sets occurred in nature (e.g. stratigraphic
unit with cross-beds joints). Note that other termination types can be collected in the field for subsequent
analysis of fracture length; therefore their definition depend on the size of the assumed window and is not
directly linked the structural character of the rock mass, but they are indicators of censoring biases
introduced by the sampling method.
Figure 7 – Example of fracture termination types for characterisation of fracture connectivity and structural
character of the rock mass (top) and characterisation of censoring biases for fracture length analysis


CONCLUSIONS
The DFN approach takes full advantage of fracture data collected from mapping of exposed
surfaces and boreholes, including also digital photogrammetry and laser scanning techniques. However,
there are important aspects engineers should be consider when collecting fracture data for DFN analysis.
Limitations inherently introduced by the sampling methods being adopted represent important aspects that
should drive the collection of discontinuities data for DFN analysis. Accordingly, the authors recommend
the introduction of a specific set of appropriate DFN guidelines to bridge the gap between the current, very
practical, ISRM methods for mapping of discontinuities in the field and the data requirements imposed by
the use of new DFN technologies.
REFERENCES
Dershowitz, W. & Einstein, H.H. (1988). Characterizing rock joint geometry with joint system models.
Rock Mech. and Rock Eng. Vol. 21. pp. 21-51.
Dershowitz, W.S. & Herda, H.H. (1992). Interpretation of fracture spacing and intensity. Proc. 33rd U.S.
Rock Mech. Symp., Santa Fe, NM.
Elmo, D., Rogers S., Stead, D. & Eberhardt, E. (2014). A Discrete Fracture Network approach to
characterise rock mass fragmentation and implications for geomechanical upscaling. Transactions
of the Australian Inst. Mining Metallurgy. (in Press).
Elmouttie, M.K. & Poropat, G.V. (2011). Uncertainty propagation in structural modeling. Proc. Int. Symp.
on Rock Slope Stability in Open Pit Mining and Civil Engineering, Vancouver, Canada.
Hadjigeorgiou, J. (2012). Where do the data come from? Proc. of Deep Mining 2012. Edited by Y. Potvin.
Australian Centre for Geomechanics, Perth, ISBN 978-0-9806154-8-7.
Harrison, J.P. (2012). Rock Engineering, uncertainty and Eurocode 7: implications for rock mass
characterisation. Proc. of MIR 2012-XIV, Faculty of Engineering, Turin University, Italy.
ISRM (1981). International Society Rock Mechanics suggested methods for rock characterization testing
and monitoring. Brown, E T Editor Imprint Oxford: Pergamon. 211 pp.
Mauldon, M. (1998). Estimating mean fracture trace length and density from observations in convex
windows. Rock Mech. Rock Eng. Vol. 31. pp. 201-216.
Rogers, S., Elmo, D., Webb, G. & Catalan, A. (2014). Volumetric fracture intensity measurement for
improved rock mass characterisation and fragmentation assessment in block caving operations.
Rock Mech. Rock Engng. DOI 10.1007/s00603-014-0592-y.
Staub, I., Fredriksson, A. & Outters, N. (2002). Strategy for a rock mechanics site descriptive model
development and testing of the theoretical approach. SKB Report R-02-02. Golder Associates AB.
SKB.
Uzielli, M. (2008). Statistical data of geotechnical data. Geotechnical and Geophysical Site
Characterisation. London. Taylor and Francis. pp 173-194.
Wang, X. (2006). Stereological Interpretation of Rock Fracture Traces on Borehole Walls and other
Cylindrical Surfaces. Ph.D. Thesis, Virginia Polytechnic Institute and State University.
Zhang, L. & Einstein, H.H. (2000). Estimating the intensity of rock discontinuities. Int. J. Rock Mech. and
Min. Sci. Vol. 37. pp. 819-837.


Guidelines For The Quantitative Description of Discontinuities For Use in Discrete Fracture Network Modelling

Uploaded by

Document Informationclick to expand document information

Copyright:

Available Formats

Guidelines For The Quantitative Description of Discontinuities For Use in Discrete Fracture Network Modelling

Uploaded by

Document Information

Original Title

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

Guidelines For The Quantitative Description of Discontinuities For Use in Discrete Fracture Network Modelling

Uploaded by

Copyright:

Available Formats

Innovations in Applied and Theoretical Rock Mechanics Editors – Hassani, Hadjigeorgiou, Archibald

Downloaded from http://onepetro.org/isrmcongress/proceedings-pdf/CONGRESS13/All-CONGRESS13/ISRM-13CONGRESS-2015-246/1473531/isrm-13congress-2015-246.pdf by University of British Columbia user on 09 December 2021