The Warm Blanket of Geotechnical Databases

and the Reality of Rock Mechanics

Julian Watson1(B) , Antonio Verduzco Mendez2 , and Patricia Robertson2

1 Mining One Pty Ltd, Bellingham, USA
[email protected]
2 Globe, USA

Abstract. The limited success of simplified slope stability assessment tools and
concerns related to slope performance resulted in a review of an open pit mine
located in South Western United States. 3D numerical was used to assess the LOM
design which led to improved slope stability, while maintaining key safety and
production targets.
An advanced inelastic strain-softening constitutive model (IUCM) was used
with FLAC3D (finite difference) and SlopeX software to assess slope stability
of the LOM design. The weak nature of the rock mass necessitated that time-
dependent behaviour was accounted for, which typically cannot be considered by
simplified slope stability assessment approaches.
The limited availability of a well-developed geotechnical database required
that model inputs (material properties) were primarily derived through rigorous
back analysis of slope behaviour, as recorded by slope monitoring equipment.
Significantly, subsequent slope performance of the LOM compares favourably
with original forecast results derived from the advanced numerical software and
IUCM.
This paper demonstrates that the derivation of material properties along with
the simulation of complex rock mass behaviour can be successfully achieved
despite a limited geotechnical database. This underlines the importance of apply-
ing the correct assessment tools and rock mass constitutive models in order to
produce safe, reliable open pit mine designs.

1 Introduction
1.1 Statistical Versus Engineering Analysis
Rock masses are naturally occurring substances that are typically variable and com-
plex in their makeup. It is this feature that geostatistics and resource modeling of min-
eral deposits attempts to understand through the development of a geological drillhole
database with sufficient quality and quantity. It would seem reasonable, therefore, that
the same method could be used when characterising variability of rock mass quality
for geotechnical assessment purposes. Considerable resources are regularly allocated to
develop geotechnical databases through approaches such as field work (core logging and
mapping) and laboratory/index testing. The resulting database is often used to classify
rock mass quality (Q, RMR, GSI etc.) and/or to develop block models and synthetic

© The Author(s) 2023

S. Javankhoshdel and Y. Abolfazlzadeh (Eds.): TMIC 2022, AHE 13, pp. 103–117, 2023.
https://doi.org/10.2991/978-94-6463-104-3_11
104 J. Watson et al.

Fig. 1. Statistical Analysis Compared to Engineering Analysis

rock masses on which geotechnical assessments are based. However, it takes consider-
able time and expense to develop a geotechnical database, meaning data quantity and
overall quality can be adversely affected if such resources are not available. Faced with
such dilemmas, many practitioners advocate a ‘simple’ geotechnical assessment tool be
used. However, doing this inadvertently assumes a false equivalence between rock mass
variability (statistical analysis) and rock mass response to loading (engineering analysis)
(see Fig. 1). In reality, these analyses approaches are vastly different and should not be
confused nor conflated. A well-developed geotechnical database cannot make up for
the shortcomings of an overly simplified engineering assessment tool. But is the reverse
true? Can an appropriate engineering assessment tool make up for the shortcomings of
a limited geotechnical database? This paper will show, through a slope stability case
study, that this can be achieved.

1.2 Slope Assessment Methods

Slope stability assessments are commonplace throughout the mining industry and are
largely used to evaluate safety and economic risk. The different assessment approaches
largely fall into three categories; empirical methods; limit equilibrium methods; and
numerical methods.
Empirical methods (formal and informal) have been applied in mine design for
many decades. The reliability of empirical methods is reliant on the database of cases
(on which they are based) being comparable with the specific condition being assessed.
Furthermore, such methods cannot take into consideration site-specific aspects, such as
unusual 3D mine geometries or ground conditions. It is for this reason that many slope
assessments are now completed using limit equilibrium or numerical methods, which
can take many site-specific aspects into account.
 The Warm Blanket of Geotechnical Databases 105

The application of limit equilibrium and numerical methods for slope stability assess-
ments is widespread. However, there are considerable and well-documented differences
between these methods. Aspects such as equilibrium, internal slope stresses, rock mass
deformation, slope failure, and kinematics are assessed in considerably different ways.
Significantly, these differences can affect assessment outputs beyond factor of safety
(FoS) results (Chiwaye, 2010). According to Lorig and Varona (2004), the main con-
tributions of numerical models in rock slope stability analyses can be summarized
as:

• Numerical models can automatically extrapolate potential failure mechanism(s) and

the most probable failure surface (or shear zone).
• Numerical models can implicitly or/and explicitly incorporate significant geological
features (such as faults, weathered zones, etc.) and groundwater conditions providing
more realistic analysis than classical analytical models in which the conditions are
frequently oversimplified leading to conservative solutions.
• Numerical analyses can help to explain in more detailed manner the physical behaviour
of slope instability.
• Numerical methods can test, in a relatively short time, multiple topographical
geological and failure mechanism situations and propose different design options.

1.3 Numerical Methods and Constitutive Models

While numerical methods are regularly used to assess rock slope stability, many geotech-
nical practitioners overlook several key aspects; the numerical code and constitutive
model used. Consequently, many assume that all numerical methods generate the same
or similar results. This is incorrect. As is discussed below, selecting the correct numerical
code and constitutive model is critical if a reliable slope stability assessment is to be
produced.

Numerical Codes
Vakili and Sandy (2014) provide a detailed discussion on common numerical modelling
methods used in the mining industry. The most widely applied numerical modelling
methods in the mining industry are described below. It is intended that these be general,
non-technical descriptions, intended to enable key concepts to be understood by a wider
audience in the mining industry. Consequently, it should be noted that some of the terms
used might not be accurate from a purely technical perspective. It is important to also note
that all analysis software and tools have their own benefits and limitations. Furthermore,
there is no single tool or software that is applicable to every single scenario. Each method
is described in terms of its predictive power. Predictive power here is measured as the
ability of a model to forecast displacement, stress, extent of fracturing, and extent of
failure induced as a result of a single excavation; in a broad sense, its ability to predict
or replicate the observed response to mining:

• Elastic Methods: This method assumes a linear elastic mechanical response for the
rock material. Consequently, its predictive power significantly reduces with more
106 J. Watson et al.

complex conditions (such as anisotropic, highly stressed, and squeezing rock). Inter-
pretation of results is achieved through indirect and empirically-derived failure cri-
teria. Consequently, reliability of this method outside its predictive limits depends
greatly on the adopted failure or interpretation criteria and its calibration efforts. The
most widely used ‘elastic’ software codes are MAP3D and EXAMINE3D.
• Inelastic Time-Independent (Implicit) Methods: The inelastic time-independent
(implicit) method can account for both elastic and non-elastic mechanical responses
of rock material. However, this method uses an implicit (time-independent) numerical
technique which finds a solution by solving an equation involving both the current state
of the system and its future state. This means that these codes are usually unable to
follow the development of rock mass failure, including progressive damage processes
in a system, nor extreme material non-linearity such as softening. The predictive power
of this method is usually better in small deformation/strain conditions, and should
always be accompanied by appropriate calibration and manual material changes when
used outside its predictive limit. The chosen failure criterion can also greatly influence
the predictive power of this method. Abaqus/Implicit and RS2/RS3 are the most
commonly used software codes that employ the inelastic time-independent method.
• Inelastic Time-Dependent (Strain Dependent) (Explicit) Methods: This method
estimates the future state of a model from its current state in an automatic time step-
ping manner, which enables the development of a failure, including non-linear material
changes due to excessive damage (i.e., new phases of stress and strain). This method
can accommodate a wide range of rock mass conditions, including high-stress and
high-deformation situations. When used with an appropriate constitutive model (fail-
ure criterion) and carefully selected input parameters and pre-mining stress state, this
method offers a reasonable level of accuracy. Consequently, often only minor cali-
bration efforts are required when monitoring results become available, meaning that
it is more a case of validation than calibration. The outputs from this approach are
often directly comparable to outputs from real-life monitoring data such as displace-
ment, strain, and stress. Abaqus/Explicit and FLAC3D are the most commonly used
continuum software codes that feature this method, while ELFEN, 3DEC (if used
in fully-discontinuum mode), or PFC are the more commonly used discontinuum
examples.

Constitutive Models
A constitutive model is a mathematical representation of a material that describes the full
mechanical behaviour from none to complete damage. Depending on its complexity, a
constitutive model depends on a number of parameters that usually represent some
mechanical characteristics of the material being modelled. Unlike discontinuum codes,
continuum numerical methods depend heavily on constitutive models to represent how a
rock mass responds to loading. Much like numerical methods, the subject of constitutive
models is wide but usually falls into one of three categories (Vakili, 2017):

• Linear Elastic Model: The linear elastic model assumes that rock behaves elastically
under applied loading, meaning that it will return to its initial shape and size when
any disturbing forces are removed. In reality, the rock may remain elastic only up to
 The Warm Blanket of Geotechnical Databases 107

about half to two-thirds of its peak strength. Consequently, this model increasingly
loses its predictive power as the level of loading and failure increases in the rock mass.
Elastic models require only a few input parameters, and there’s often a standard and
easy guideline for selection of these inputs.
• Strain-Softening-Mohr-Coulomb Model: The Mohr–Coulomb peak strength cri-
terion and associated constitutive models have gained wide acceptance and applica-
tion in the field of geotechnical engineering. This criterion is obtained from two key
parameters; cohesion (c) and friction angle (F). Although, more advanced numerical
assessments often couples the Mohr–Coulomb failure criterion with a strain-softening
post-peak model to simulate the full stress–strain behaviour of the rock mass. (Note
that there are other, less well known variations also by some practitioners.)
• Proprietary Model: These models include proprietary algorithms (i.e., owned / com-
mercialised intellectual property) and processes that are used to solve the full stress–
strain behaviour within the numerical model. While the capability and accuracy of
these models is often promoted by their developer, several fundamental issues arise,
including lack of transparency (such as technical details, theories, and limitations)
and difficulty to independently review (due to limited information provided or access
to the material model). Examples of advanced rock mechanics constitutive models
include, but are not limited to LR2 (Levkovitch, Reusch, & Beck, 2010) and IMASS
(Itasca, 2022).

Improved Unified Constitutive Model (IUCM)

The Improved Unified Constitutive Model (IUCM) was developed with the intention to
address the limitations of the current modelling practices described above in this paper.
This model gathers the most notable and widely accepted previous research work in the
area of rock mechanics and integrates it into a unified constitutive model that can better
and more accurately predict the stress–strain relationships in a continuum model. The
IUCM was developed after extensive calibration against well-established case histories
and independent field applications, including weak and strong rock masses (Vakili, 2016;
Vakili, 2017). Such case histories include (but are not limited to):

• Triaxial test results:

– Various rock mass failure mechanisms which are dependent on different confine-
ment levels (Tiwari & Rao, 2006).
– Analytical relationships proposed by Hoek and Brown (1980), and laboratory test
results outlined by Donath (1972), and McLamore and Gray (1967).

• Failure mechanisms in massive, brittle rock associated with the Mine-by-Experiment

project (Vakili, 2016).
• Complex buckling failure mechanism and effects of rock strength anisotropy associ-
ated with vertical shafts (Vakili, Sandy, & Albrecht, 2012; Vakili, Albrecht, & Sandy,
2014; Watson, Vakili, & Jakubowski, 2015).
• Back analysis of observed hanging wall damage in a deep underground open stope
(Vakili, Albrecht, & Sandy, 2014).
108 J. Watson et al.

• Pillar stability assessment of a massive brittle rock mass (Vakili, 2016).

• Open pit mine slope stability assessments (Sainsbury, Vakili, Lucas, & Hutchison,
2016; Hutchison, Chambers, Gannon, & Oko-oboh, 2017; Roach & Johnston, 2020;
Vakili, Watson, Abedian, & Styles, 2020; Ford, Lucas, & Vakili, 2020; Lucas, Vakili, &
Hutchison, 2020; Hutchison, Morrison, & Lucas, 2020).

It should be noted that the IUCM model calibration results for open pit applications
compare favourably with more traditional elastic and coupled strain-softening-Mohr-
Coulomb (SSMC) models. The linear nature of the residual envelope in the IUCM with

Fig. 2. Conceptual Representation of the Post Peak Rock Mass Behaviour as Represented in the
IUCM; (a) Principal Stress Space; and (b) Stress-Strain Space (Vakili, 2016)
 The Warm Blanket of Geotechnical Databases 109

Fig. 3. Conceptual Description of the IUCM

regards to the peak Hoek-Brown envelope, replicates cohesion and friction softening
at lower confinement levels and cohesion-softening and friction-hardening at high con-
finement levels (Fig. 2). This allows progressive failure to occur near the boundary of
excavations, while limiting the propagation of yield or plasticity zones away from exca-
vation boundaries (Vakili, 2016). A conceptual description of the IUCM is presented in
Fig. 3. Where possible, the development of new theories or techniques was avoided when
devising the IUCM, with a focus on well-accepted and widely applied rock mechanics
techniques and theories. Furthermore, the inputs, limitations, and form of the IUCM are
transparent and without ambiguities often associated with proprietary and commercially
sensitive aspects of some models.

2 Open Pit Case Study

A technical review of an open pit mine located in South Western United States was under-
taken after slope performance concerns were identified by site personnel. This included
large tension cracks developing behind the open pit crest, along with large displacement
occurring over multiple benches (particularly where deposit-scale faults coalesce with
the open pit design surface and each other). Consequently, concerns were raised that
the LOM design may lead to large-scale slope instability that results in unacceptably
high risk associated with safety, the environment, and mine production. It is noted that
the rock mass was composed of a number of pit-scale faults, and was characterised by
several weak rock units. Furthermore, it was recognised that complex ground conditions
and time-dependent slope behaviour demanded a more advanced assessment approach to
reliably forecast slope stability and effectively manage associated hazards. Significantly,
the geotechnical database was limited to some mapping data and material properties used
110 J. Watson et al.

in a previous 2D limit equilibrium assessment. A high-quality slope monitoring database

of prism and radar data was available.

2.1 Slope Stability Assessment Approach

Given the limited geotechnical database, it was decided that a two-phase approach would
be used; back analysis of slope failures using slope monitoring data to derive material
properties, followed by forward analysis of the LOM design. Both phases of the study
applied the SlopeX plugin (Vakili, Watson, Abedian, & Styles, 2020) and FLAC3D
software (Itasca, 2012), along with the IUCM.
It was assumed that the adopted vertical in situ stress is lithostatic (i.e., based on
the density of the overlying rock) and that the major horizontal in situ stress magnitude
equals the vertical stress magnitude, oriented east-west.
Site personnel initially indicated that ground water was located beneath the LOM
open pit design. Consequently, the models did not include a phreatic surface (i.e.,
were run ‘dry’). (It should be noted that subsequent information revealed several faults
contained ground water, suggesting results may be ‘best case’ in affected areas.)

2.2 Phase 1: Model Calibration

A calibration model was constructed using available data from field mapping, prism, and
scanning monitoring records, previous 2D limit equilibrium assessment assumptions,
photographs, and conversations with site personnel. The absence of core logging and
laboratory testing data should be noted. Therefore, an iterative approach was applied:

Fig. 4. Example of Phase 1 Results. Displacement of Upper Slope Compared with Prism Data
(Blue = Prism Data, Red = Calibration Model)
 The Warm Blanket of Geotechnical Databases 111

Fig. 5. Example of Phase 1 Results from July 2020. Displacement of Lower Slope Compared
with Prism Data (Blue = Prism Data, Red = Calibration Model)

• Step 1: Apply rock mass and fault material properties used in previous 2D limit
equilibrium assessments and exposure mapping. All key geological features (such as
lithology and fault wireframes) were included in the numerical models. Results were
compared to historical rock mass response to mining (i.e., slope monitoring data and
photographs).
• Step 2: Access monitoring data records from the slope, selecting data points from
representative areas of the slope that have undergone various ranges of displacement
(i.e., small to large).
• Step 3: Based on key monitoring prism data and model runs, modify initial material
properties to match monitoring data.
• Step 4: Re-run numerical models iterations until a close fit with monitoring data results
and overall slope behaviour is identified.

2.3 Phase 1: Model Results

Examples of results from the Phase 1 calibration model are presented in Fig. 4 (upper
slope) and Fig. 5 (lower slope). It can be seen that displacement results compare
favourably with slope prism monitoring data for the same mining period. Model sections
through the West Lobe slope suggest that shallow bench-scale displacement and deeper
shear strength results compare favourably with slope behaviour observed in the field.
Results indicated the rock mass to be significantly weaker than was thought to be
the case prior to commencement of mining in 2017 and previous 2D limit equilib-
rium assessments (Table 1). Therefore, initial material properties were downgraded by
approximately 30% during the calibration phase. This was commensurate with historical
time-dependent ‘creeping’ behaviour recorded on the slope.
112 J. Watson et al.

Table 1. Material Properties Developed from Calibration Model

Geotechnical GSI Sigci mi V coh ϕ Ei D Residual

Domain (MPa) (kPa) (°) (MPa) Friction
Angle (°)
Initial Pinal Schist 40 35 8 0.3 230 51 12,500 0 45
Schultz 55 100 29 0.2 1200 68 2,465,600 0 45
Granite
Apache 60 70 25 0.2 850 65 2,842,700 0 45
Leap Dacite
Apache 35 25 4 0.2 300 45 725,200 0 45
Leap
Vitrophyre
Faults(1) 25 25 2 0.2 20 25 725,200 0 45
Final Pinal Schist 30 20 8 0.3 230 51 543,900 0 44
Schultz 40 30 29 0.2 1200 68 739,700 0 44
Granite
Apache 45 21 25 0.2 850 65 852,800 0 44
Leap Dacite
Apache 25 8 4 0.2 300 45 217,500 0 44
Leap
Vitrophyre
Faults(1) 20 8 2 0.2 20 25 217,500 0 30
Joints 5 2 25 0.2 20 25 435,100 0 29
(1) Fault Zone assumed to be ~2–3 m wide

2.4 Phase 2: Forecast of LOM Design Results

Using the material properties derived from the Phase 1 calibration model, a forward
analysis model was run to assess slope behaviour of the LOM design (finishing in
February 2022). The focus being to confirm that the LOM can be achieved. Results of
model velocity and FoS are presented in Fig. 6 and Fig. 7 respectively.

2.5 Comparison of Forecast Model with Final Slope Behaviour

Results from the Phase 2 forecast model suggested that slope stability of the LOM open
pit design met industry standard acceptability criteria:

• Overall slope angle-scale slopes were predicted to be stable throughout the LOM
design, with a FoS > 1.3. This is in line with the acceptability criteria for this scale of
slope with moderate consequence for failure (Read & Stacey, 2009). Model velocity
results also met or exceeded recommended thresholds of 1e-7 m per timestep (Lorig &
Varona, 2000).
• Inter-ramp angle-scale slopes were also predicted to be stable throughout the LOM
design, with a FoS > 1.2. While this also met or exceeded the acceptability criteria
 The Warm Blanket of Geotechnical Databases 113

Fig. 6. Forecast Model Velocity Results of LOM (February 2022) Open Pit Design, Suggesting
Shallow, Bench-Scale Failure (i.e., No Deep-Seated Slope Failure)

Fig. 7. Forecast Model FoS Results of LOM (February 2022) Open Pit Design, Suggesting
Shallow, Bench-Scale Failure (i.e., No Deep-Seated Slope Failure)

of > 1.2 FoS for slopes with moderate consequence for failure (Read & Stacey,
2009), results suggested retrograde surficial bench-scale degradation would likely
occur (particularly where deposit-scale faults coalesce with open pit design and each
other).

The Phase 2 model forecast results were compared with actual LOM prism moni-
toring array records at completion of mining (i.e., LoM) in February 2022. As was pre-
dicted, overall slope stability was maintained, despite considerable retrograde surficial
bench-scale degradation. Furthermore, the forecasted extent and location of degradation
(displacement) compared closely with the actual data as can be seen in Fig. 8 and Fig. 9.
114 J. Watson et al.

Fig. 8. Forecast Model Displacement Results Projected onto The LOM (February 2022) Open
Pit Design, Along with Actual Prism Data Collected Over the LOM (Spheres)

Fig. 9. As-Built LOM Slope at Completion, February 2022 (Looking North)

 The Warm Blanket of Geotechnical Databases 115

2.6 Conclusion

This paper shows that an appropriate engineering assessment tool can make up for the
shortcomings of a limited geotechnical database. While a well-developed geotechni-
cal database is always preferable, it is demonstrably not necessary (i.e., statistical and
engineering assessments are independent).
Other benefits of using an appropriate engineering assessment tool, as shown through
this case study, include (but are not limited to):

• Confidence that the overall slope stability would be maintained throughout the LOM
design.
• Confirmed the need for rigorous slope monitoring practices (including prism and radar
monitoring arrays, along with regular visual inspections).
• Provide a “road map” for when and where retrograde surficial bench-scale degrada-
tion would likely occur throughout the LOM (as evidenced by the position of prism
locations and forecast slope behaviour in Fig. 8).

2.7 Limitations and Uncertainties

There are a number of limitations and uncertainties that should be considered when
assessing numerical modelling results of open pit slope stability. A summary of these
limitations include:

• The numerical modelling often does not account for any dynamic effects associated
with blasting, nor the effects of rock mass degradation caused by weathering and
erosion.
• The confidence in the condition and location of modelled fault structures (which can
greatly influence results).

References
Chiwaye, H. T. (2010). A comparison of the limit equilibrium and numerical modeling approaches
to risk analysis for open pit mine slopes. MSc Thesis at University of the Witwatersrand.
(Unpub).
Donath, F. A. (1972). Effects of cohesion and granularity on deformational behaviour of anisotropic
rock. Geological Society of America, 135, 95-128.
Ford, A., Lucas, D. S., & Vakili, A. (2020). Validation of improved unified constitutive model
(IUCM) for open pit applications. Perth: Slope Stablity 2020 (in publication).
Hoek, E., & Brown, E. (1980). Underground Excavations in Rock. London lnstn. Min. Metall.
Hutchison, B. J., Chambers, J., Gannon, S., & Oko-oboh, D. (2017). Case study of joint structure
and the design and performance of open pit walls at the Kanmantoo Copper Mine. Hobart,
Australia: Proceeding of Tenth International Mining Geology Conference.
Hutchison, B. J., Morrison, A. T., & Lucas, D. S. (2020). Steep wall mining – experience in the
management of rockfall hazards at Kanmantoo Copper Mine. Perth: Slope Stablity 2020 (in
publication).
116 J. Watson et al.

Itasca. (2012). FLAC3D, Fast Lagrangian Analysis of Continua in 3 Dimensions (V5). Itasca
Consulting Group.
Itasca. (2022, September 6). IMASS Model. Retrieved from Itasca Consulting Group Inc.: http://
docs.itascacg.com/flac3d700/common/models/imass/doc/modelimass.html
Levkovitch, V., Reusch, F., & Beck, D. (2010). Application of a Non-linear Confinement Sensitive
Constitutive Model to Mine Scale Simulations Subject to Varying Levels of Confining Stress.
ISRM International Symposium. Lausanne, Switzerland: EUROCK 2010.
Lorig, L. J., & Verona, P. (2004). Numerical analysis in rock slopes. In D. C. Wyllie, & C. W. May
(Eds.), Rock Slope Engineering - Civil and Mining (4th Edition) (p. Ch 10). London: Spoon
Spress.
Lorig, L., & Varona, P. (2000). Practical slope-stability analysis using finite-difference codes. In W.
A. Hustrulid, M. K. McCarter, & D. J. Van Zyl, Slope Stability in Surface Mining (pp. 115-124).
Society for Mining and Metallurgy, and Exploration, Inc: Littleton.
Lucas, D. S., Vakili, A., & Hutchison, B. J. (2020). Slope design and control of rock fall hazards
in a challenging structural setting at the Kanmantoo copper mine, South Australia. Perth: Slope
Stablity 2020 (in publication).
McLamore, R., & Gray, A. E. (1967). The mechanical behaviour of anisotropic of sedimentary
rocks. The American Journal of Mechanical Engineers, 89(1), 62-73.
Read, J., & Stacey, P. F. (2009). Guidelines for Open Pit Design. Melbourne: CSIRO Publishing.
Roach, B., & Johnston, A. R. (2020). Engineering geology investigation and numerical modelling
design of the ramp 12 highwall. Perth: Slope Stablity 2020 (in publication).
Sainsbury, D. P., Vakili, A., Lucas, D. S., & Hutchison, B. J. (2016). Three-dimensional numerical
modelling of potentially structurally controlled failure mechanisms at the Kanmantoo open pit.
Perth: Proceedings of the First Asia Pacific Slope Stability in Mining Conference, Australian
Centre for Geomechanics.
Tiwari, R. P., & Rao, K. S. (2006). Post failure behavior of a rock mass under the influence of
triaxial and true triaxial confinement. Engineering Geology 84(3), 112-129.
Vakili, A. (2016). An improved unified constitutive model for rock material and guidelines for its
application in numerical modelling. Computers and Geotechnics, 80, 261-282.
Vakili, A. (2017). The improved unified constitutive model: a fine-tuned material model tailored for
more challenging geotechnical conditions. Deep Mining 2017: Eight International Conference
on Deep and High Stress Mining, J. Wesseloo (Ed.). Australian Centre for Geomechanics.
Vakili, A., & Sandy, M. (2014). Geotechnical analysis – not just another box to be ticked. Bulletin,
36-40.
Vakili, A., Albrecht, J., & Sandy, M. (2014). Rock strength anisotropy and its importance in
underground geotechnical design. Proceedings of the third Australasian ground control mining
conference (AusRoc 2014) (pp. 167-80). Melbourne: The Australasian Institute of Mining and
Metallurgy.
Vakili, A., Sandy, M., & Albrecht, J. (2012). Interpretation of non-linear numerical models in
geomechanics - a case study in the application of numerical modelling for raise borded shaft
design in a highly stressed and foliated rock mass. MassMin 2012: proceedings of the 6th
international conference and exhibition on mass mining. Sudbury: MassMin.
 The Warm Blanket of Geotechnical Databases 117

Vakili, A., Watson, J. M., Abedian, B., & Styles, T. (2020). An introduction to SlopeX – a Plug-in
to simplify and fast-track advanced numerical modelling for open pit applications. Perth: Slope
Stablity 2020 (in publication).
Watson, J. M., Vakili, A., & Jakubowski, M. (2015). Rock Strength Anisotropy in high stress con-
ditions: a case study for application to shaft stability assessments. Wroclaw: Studia Geotechnica
et Mechanica.

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-
NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/),
which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any
medium or format, as long as you give appropriate credit to the original author(s) and the source,
provide a link to the Creative Commons license and indicate if changes were made.
The images or other third party material in this chapter are included in the chapter’s Creative
Commons license, unless indicated otherwise in a credit line to the material. If material is not
included in the chapter’s Creative Commons license and your intended use is not permitted by
statutory regulation or exceeds the permitted use, you will need to obtain permission directly from
the copyright holder.