See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/341313162

3D limit equilibrium method for rock slope stability analysis using generalised
anisotropic material model

Conference Paper · May 2020

DOI: 10.36487/ACG_repo/2025_46

CITATIONS READS

10 2,240

2 authors:

Sharan Kumar Nagendran Mohd Ashraf BIN Mohamad Ismail

Malaysian Public Works Department (PWD) Universiti Sains Malaysia
22 PUBLICATIONS 228 CITATIONS 111 PUBLICATIONS 624 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Sharan Kumar Nagendran on 12 May 2020.

The user has requested enhancement of the downloaded file.

 Slope Stability 2020 - PM Dight (ed.)
© Australian Centre for Geomechanics, Perth, ISBN 978-0-9876389-7-7
https://papers.acg.uwa.edu.au/p/2025_46_Ismail/

3D limit equilibrium method for rock slope stability analysis

using generalised anisotropic material model

NS Kumar Universiti Sains Malaysia, Malaysia & Centre of Excellence for Engineering and Technology
(CREaTE) JKR, Malaysia
MAM Ismail Universiti Sains Malaysia, Malaysia

Abstract
Rock anisotropy is a well-known phenomenon relating to the heterogeneity of rock mass. Nevertheless, its
influence in geotechnical design, especially in rock engineering, is often ignored. Slope with certain modes of
failure can be evaluated conventionally as well as numerically. For this study, the rock slope assessment was
conducted numerically using 2D and 3D limit equilibrium method (LEM) utilising the Slide program by
Rocscience. The fundamental roles of the discontinuities present in the study area were evaluated to study
their influence on slope stability. Anisotropic material model was incorporated in the LEM analysis to
investigate the presence of discontinuities. The measurement of discontinuity orientation in the rock slope by
traditional scanline survey is time-consuming and challenging due to accessibility issues. Structure from
Motion (SfM) photogrammetry using unmanned aerial vehicle (UAV) allows a quick and cost-effective way to
do survey mapping for geotechnical assessment on rock slope compared to terrestrial laser scanner. Dense
point cloud is exported to the CloudCompare tool for geological plane extraction. The stability of the rock
slope was evaluated using the deterministic 3D and 2D LEM using the geometry of the 3D rock slope system.
In this study, the anisotropic material model was utilised to examine the Factor of Safety (FoS) results.
Generalised anisotropic material model was used for incorporating the generalised Hoek–Brown criterion
(rock mass) and Barton–Bandis criterion (weak joint). The rock mass and shear strength parameter for
numerical analysis were determined via destructive and non-destructive tests such as uniaxial compressive
strength, Schmidt hardness and joint roughness coefficient estimation using Barton comb. Mean dip/dip
direction obtained was used as an input for the value of the anisotropic plane where it causes a weakness in
the strength of the rock slope. The results of FoS shows that rock slope without anisotropy model is stable and
analysis using anisotropic material model predicts that the slope may fail. 3D slope stability analysis was able
to identify the weakest spot easily rather than making an assumption based on the results of 2D slope stability
assessment which might represent the whole rock slope. 3D rock slope stability assessment proves to be a
very cost-effective method for remedial work whereas in 2D stability assessment, wrong cut-sections may
provide inaccurate FoS. This study presents the approach of using anisotropic material model utilising basic
rock testing and field observation data to analyse the rock slope stability.
Keywords: anisotropy, slope stability, limit equilibrium analysis, generalised Hoek–Brown, Barton–Bandis

1 Introduction
The assessment of the stability of natural rock slopes in hilly terrains areas becomes very necessary for
implementing safe construction practices. Instability is determined by the occurrence of pre-existing
geological features such as joints, bedding planes or rock slopes failures. Failure related to the geological
features will normally be in the form of plane sliding, wedge sliding or toppling. Mostly, failure occurs in a
rock slope because of one or a combination of these three mechanisms. There are many methods available
for evaluating the stability of slope in term of the safety factors, such as limit equilibrium method (LEM), limit
analysis method, finite element method and finite difference method. However, there is a need to emphasise
the differences in terms of the advantages and parameters of each methods. LEM is a method of essentially
dividing a soil/rock mass above the surface of the slip into a finite number of slices where the slices can be

Slope Stability 2020 715

3D limit equilibrium method for rock slope stability analysis using generalised NS Kumar and MAM Ismail
anisotropic material model

segmented vertically or horizontally (Cheng et al. 2008; Agam et al. 2016). Numerical simulation methods
have gained much attention due to their numerous benefits over LEM. Conventional numerical modelling
based on different methods for analysing slope stability in situations where discrete geological structures did
not totally control the failure mechanism were not efficient.
It is a general conception that 2D slope stability analysis is too conventional and inadequate in comparison
with 3D (Cheng et al. 2005; Leong & Rahardjo 2012; Nian et al. 2012). Owing to these limitations of apparent
instability originating from 2D analysis, comprehensive and practical 3D models were developed. Assessment
of 2D slope stability sometimes leads to over-simplification of the issue and may result in incorrect
assessment of the potential failure mechanism. Several researchers have reported that 2D models cannot
solve a 3D problem in many situations (Cala 2007; Huang & Tsai 2000; Nian et al. 2012). Rock mass is largely
discontinuous and anisotropic by nature, and this makes the rock a difficult material to be represented
mathematically in numerical modelling. Nevertheless, due to the strong anisotropic properties between the
normal direction and tangential direction of the bedding joints, the assumption cannot be appropriate for
the layered broken rock slope. Amadei (1996) stressed the importance of joint rock mass anisotropy.
Different methods have examined anisotropic properties such as elastic modulus and permeability (Gonzaga
et al. 2008; Nunes 2002; Snow 1969).
Uniaxial compressive strength (UCS) test, triaxial compressive strength test, point load test, and Schmidt
hammer test are the laboratory test conducted to evaluate rock strength. A stone of 54 mm in diameter is
cut from the core and placed under the pair of pistons to be compressed by axial loads. Rock mechanical
laboratory testing on core samples is the most accurate method for calculating rock strength. Deere & Miller
(1966) also suggested the use of the Schmidt rebound hammer to measure joint compressive strength.
Surface roughness is said to increase the surface's shear strength and this improvement in strength is
extremely important in terms of rock excavation stability. Using the Barton comb profilometer, the surface
roughness is generated (Kim 2016; Özvan et al. 2014).
Photogrammetry technique has become easier and user-friendly following the recent development of the
Structure from Motion (SfM) system and related technologies. Westoby et al. (2012) are providing an
introduction to this methodology while Salvini et al. (2017) and Francioni et al. (2018) also recently discussed
their applications. SfM is based in multiple overlapping images on a highly redundant bundle adjustment
matching features. Although this technique makes it easier and faster to build 3D models, the use of ground
control points (GCPs) is still needed for scaling and georeferencing the models (Francioni et al. 2019).
Unmanned aerial vehicle (UAV) can be used to capture digital images for photogrammetry processing as long
cost solution (Francioni et al. 2019).
This paper aims to conduct the slope stability analysis of a rock slope located in the northern peninsular of
Malaysia, that composed of a heterogeneous sedimentary rock system. The 3D analysis will lay a benchmark
for Malaysian geotechnical practitioners and the mining industry, thereby serving as the starting point for
performing 3D slope stability analysis when considering rock slope. This study consists of collecting data in
the field and laboratory testing for the rock strength parameter. Moreover, the analysis will be utilising the
parameter determined for rock slope stability study considering the anisotropy condition.

2 Overview of site and geology

Construction through karstic limestone is always a challenging task due to the uncertainties. Some part of
small 130 hills been blasted to construct a channel to drain the water into Straits of Malaca from Timah Tasoh
dam. The blasting activity results in a rock slope to be exposed facing the channel. Some parts of the slope
failed due to the presence of weak and heavily jointed rock. The exact coordinate of the study area is at
latitude 6.428794 and longitude 100.143884. The study area is located at Kampung Wai, Timah Tasoh, Perlis
(Figure 1). Most of the discontinuities are daylighting to the slope face.

716 Slope Stability 2020

 Assessment and implications for uncertainty in design

Figure 1 Geological map and site overview

It is the southernmost limestone range on the mainland of Southeast Asia. Perlis has isolated tower karst hills
as well as a long-range of calcareous hills. Perlis is a small state in northern Malaysia. Setul formations are
widely distributed in the northwest parts of Peninsular Malaysia and southern Thailand. Setul formations are
crystalline, hardly brittle, dark coloured, thick-bedded and crystal clear. Based on Figure 2, the geological
settings of the study area are identified as limestone.

Figure 2 Geological map and site overview

Slope Stability 2020 717

3D limit equilibrium method for rock slope stability analysis using generalised NS Kumar and MAM Ismail
anisotropic material model

3 Anisotropic material model

Anisotropic rock exhibit different properties in various directions. However, rock material exhibits, in most
cases, different anomalies in the physical behaviour, which cannot be neglected if a complete solution of any
given rock mechanics problem, is to be attained. A well-known behaviour pattern which is of considerable
interest in many applications is the rock anisotropy. These anisotropies are often called inherent anisotropic
and the rock corresponding to this classified as intact anisotropic rock. (Asadi 2016; Ismael et al. 2014; Kassa
& Steinar 2016). Anisotropic material models use the multi-yield plasticity formulation. It takes one failure
criterion for the matrix and virtually number of joints. The shear strength of joints or weak planes could be
described by any of the three criteria of Coulomb, Barton–Bandis and hyperbolic. Anisotropic model is a
constitutive model that describes the shear strength of an anisotropic rock mass in relation to the change in
the angle between the plane of shear, and either the predominant plane of the weakness of the rock fabric
or the predominant orientation of major structural weakness. For this study, the generalised anisotropic
model will be utilised to simulate the anisotropic condition. The generalised Hoek–Brown model will
represent the rock mass strength and the Barton–Bandis criterion will represent the weak joint strength.
The generalised Hoek–Brown criterion for jointed rock masses is introduced by Hoek et al. (2002) in which
the shape of the principal stress plot or the Mohr envelope could be adjusted by means of a variable
coefficient a as shown in Equation 1:
σ ' a
σ1 ' = σ3 ' + σci mb σ3 +s (1)
ci

where:
Mb = a reduced value of the material constant mi for the intact rock and is calculated by
Equation 2.
100
mb = mi exp (2)
28 14D

where:
s = a constant depending upon the characteristics of the rock mass as in Equation 3.

= exp (3)

Also, parameter a is a constant for the rock mass given by Equation 4:

= + − (4)

The Barton–Bandis empirical shear strength criterion proposed by Barton was used to describe the shear
strength of discontinuities. To use this criterion, the three input parameters of joint roughness coefficient
(JRC), joint compressive strength (JCS) and residual friction angle of discontinuity must be determined. Barton
and his co-workers (1971–1990) suggested the following empirical criterion for estimating the shear strength
of rock joints:
-.
= σ! " # [ %&' ()*10 + Ø1 ] (5)
σ/

Here τ, σn, JCS and JRC are shear strength, normal stress, residual friction angle, joint compressive strength and
joint roughness coefficient, respectively. For unweathered rock fractures, the residual friction angle (Ør), is
equal to base friction angle (Øb), which can be obtained from shear tests on smooth unweathered joint surfaces.

718 Slope Stability 2020

 Assessment and implications for uncertainty in design

4 Methodology

4.1 UAV photogrammetry

The slope geometry was obtained from the 3D model generated from the UAV photogrammetric method.
UAV were used to capture aerial and side images. Later, the images were imported into Agisoft Metashape
Professional program for image stitching process using SfM algorithm. UAV data acquisition was carried out
using a DJI Phantom 4 Pro quadcopter to capture the aerial and side images of the study area. Planning of
missions can be made onsite or at the office before going to the study area (Figure 3a). For this study, the
mission was made onsite considering the boundary of an area of interest to capture images to model out the
study area. Normal photogrammetry mapping must take into consideration parameters such as flight
altitude, percentage of front and side overlapping between images, and ground sampling distance.
Waypoints are calculated based on the input parameter and then the UAV mission is usually done in an
autonomous mode based on the data transferred to the aircraft. Ground Station Pro from DJI was used to
plan the flight parameters to allow the aircraft flies precisely along the pre-designed flight path. GCPs were
utilised to georeferencing the model to produce the output with absolute accuracy. Photogrammetry
software using SfM technique was able to detect the same features from several images to create the tie
point (Figure 3b). Tie points are generated automatically based on feature matching detectors and
descriptors such as scale invariant transform. Tie points generated to form three-dimensional coordinates
from aerial triangulation process will go through bundle adjustment. 3D point cloud data will be generated
after the bundle adjustment process through the photogrammetry software. 3D model of the rock slope was
used for numerical modelling in this research work. The object file (.OBJ) format of the 3D model exported
into the numerical modelling software to simulate real site condition.

Figure 3 (a) Phantom 4 Pro drone; (b) Agisoft Metashape processing shows the camera position

4.2 Discontinuity orientation extraction

The photogrammetry output which is the point cloud data were exported into the open-source tool
CloudCompare which utilises FACET plugin (Riquelme et al. 2014). The best features of the FACETS plugin is
that it has the capability to discover planar objects but also 3D points with the stereogram tool. This method
was able to acquire orientation of discontinuity (dip/dip direction) of the rock slope using an automatic
approach by extracting the discontinuity orientation of the rock slope digitally. The 3D point cloud is
systematically divided into smaller patches in a fast marching (FM) approach. The divided patches are
subsequently regrouped and as a result of which the patches have a similar size. Further, the extracted
meshes or facets can be classified into single planes and plane families based on the orientation, i.e. dip/dip
direction. The rock slope stability analysis is conducted on the stereogram generated from this process.
By selecting the outcrop portion, queries can be raised on the stereogram. Finally, the facets data can also
be exported as comma-separated-variable (CSV) ASCII file or shapefiles in order to perform subsequent

Slope Stability 2020 719

3D limit equilibrium method for rock slope stability analysis using generalised NS Kumar and MAM Ismail
anisotropic material model

supporting analysis in other software (Dewez et al. 2016). Table 1 shows the input parameters for FACET
CloudCompare processing.

Table 1 FACET CloudCompare methods input parameters

Octree level Max distance @ 99% Min point per facet Max edge length
Fast marching
algorithm 8 (0.13 m) 0.1 m 100 1m

4.3 Methods of acquiring rock strength properties

4.3.1 Uniaxial compressive strength

In this study, a few samples of rock were obtained from different sections of the rock outcrops. A rock sample
was cored into the relevant diameter and length ratio 1:2, the diameter of the cored sample was 50 mm and
the length was 100 mm, the coring procedure and standard size were fully complying with ASTM standard
D4543-85 (ASTM 2001). The two ends of the specimens were trimmed to make sure the flat-ended on the
surface. Polishing and lapping work has been carried out to ensure flat end surfaces not to exceed 0.025 mm
and side of the specimen were straight to within 0.50 mm. Tolerance of the specimen’s flatness has to be
met according to the ASTM D4543 standard. The uniaxial compression tests were conducted on Shimadzu
hydraulic universal testing machine. This machine with 1,000 kN capacity is able to apply the compressive
load at a constant strain rate of 0.5 Mpa per second (Figure 4).

Figure 4 (a) Sample tested for UCS with strain a gauge measurement; (b) Failed sample

4.3.2 Schmidt hammer

The JCS were measured using Schmidt rebound hammer in this study (Figure 5a). N-type Schmidt hammer
was used to measure the rebound value, N. The test does not involve the destruction of the sample
(non-destructive) and it can be repeated on the same sample. The index value obtained is rebound number
(N), which is an indicator of surface hardness of the sample. Total of 20 numbers was obtained and the
average value was considered as Schmidt number for that section. The N value was used to estimate the
compressive strength of the wall using JCS estimation chart showing the correlation between Schmidt
hammer rebound number, hammer orientation, UCS and Rock density by (Deere & Miller 1966). The
determination of surface hardness using Schmidt hammer value obtained following ASTM D5873-00 (ASTM
1981) standard method.

720 Slope Stability 2020

 Assessment and implications for uncertainty in design

4.3.3 Field estimation of joint roughness coefficient

JRC is probably the most commonly used approach for quantifying the roughness of a joint profile. A standard
method of JRC determination is related to field measurements, which are performed using profilometer
(Barton’s comb). The attractiveness of the JRC factor is it provides a quantitative measure of discontinuity
profile properties. JRC values vary from 0 to 20 were obtained through the comparison of the measured joint
surface geometry with the one presented in the Barton’s standard profile chart (Barton & Choubey 1977).
It is a simple and easy device to be used for recording the 2D profile of a short (30 cm) profile length on a
discontinuity surface (Figure 4b). This inexpensive device can be used easily and can be compared to standard
JRC profiles or can be traced to obtain a hard copy of the profile. Figure 5(b) shows the field application of
the device.

Figure 5 (a) Schmidt hammer reading; (b) Joint roughness measurement using Barton comb

4.4 2D and 3D limit equilibrium modelling

Discontinuities in the rock mass will be filled up with anisotropic material which does not have the same
strength as the intact rock. Their strength is weak and will induce rock slope failure. In this study, the
anisotropic plane was assigned to the rock slope to take into consideration of the rock mass strength as well
as the weak joint strength which is dipping at an angle. The critical slip surface was searched using cuckoo
search method (CS) with surface altering optimisation in the 2D analysis. It is a stochastic method which is
more suitable for rock slope, which consists of irregular and heterogeneous anisotropic plane and material
where non-circular slip surface occur. CS can locate the absolute global minimum failure surface more
effectively and efficiently (about three times faster) than the simulated annealing method (Wu 2012)
(Mohamad et al. 2014). The non-circular slip surface type was selected due to the anisotropic plane
(discontinuity). The slope stability assessment was performed using the limit equilibrium method using
GLE/Morgenstern-Price criterion since it involved horizontal and vertical force equilibrium as well as moment
equilibrium. The generalised anisotropic strength model permits to define anisotropic strength characteristic
for material using any combination of failure criteria applied across different orientations. The generalised
anisotropic strength model allows us to define anisotropic strength properties for material using any
combination of failure criteria applied over different orientations. Dip and dip direction were used to define
the anisotropy. The composite materials of the rock slope must be defined with its strength characteristic.
The A and B parameters define the angular range of the anisotropy. For this analysis, parameter A is equal
to 5 (commonly adopted in geological model to be used for slip surfaces in LEM analysis software) and B
is 10° (used for the discontinuous joints mainly for back analyses and collapsed slopes) was used to represent
the angular variation (Bar & Weekes 2017). Different failure criteria can be selected to define the strength of
the composite materials in the generalised anisotropic. For this study, rock mass strength has been expressed
through the Hoek–Brown failure criterion. Since rock mass has discontinuities, a generalised anisotropic

Slope Stability 2020 721

3D limit equilibrium method for rock slope stability analysis using generalised NS Kumar and MAM Ismail
anisotropic material model

model was used to define the orientations of the anisotropic plane in the rock slope. Bedding/weak joint was
defined by Barton–Bandis failure criterion.

5 Results and discussion

5.1 Laboratory and field observation data

Based on the laboratory test conducted using the cored sample, the summary of the results tabulated below
(Table 2). The preparation of a sample for this destructive testing needs more caution. Due to the moderately
weathered sedimentary rock, the sample tends to break during coring. All the results were compared to the
findings from other researcher’s work on limestone.

Table 2 Laboratory results using cored sample

Sample Dry density Uniaxial compressive Min Max Standard

Median Mean
number (g/cm3) strength (Mpa) values value deviation
1 2.724 50
2 2.757 69.5
3 2.746 101.4
4 2.738 80.9
50.7 101.4 67.54 73.35 18.82
5 2.723 62.84
6 2.754 56.79
7 2.735 65.58
8 2.746 99.08

The Schmidt hammer in situ strength characteristics of the rocks were determined by performing 60 Schmidt
hammer tests at different parts along the rock slope outcrops. The in situ results were plotted on the
correlated chart to obtain the JCS value (Table 3).

Table 3 Results of Schmidt hammer for joint compressive strength estimation

Section Schmidt Average Joint compressive

hammer (N) strength (MPa)
1 35.2
2 28.0
32.3 52
3 32.1
4 33.9

JRC results from the Barton comb summarised based on the data collected at three sections section of the
rock slope outcrops. A total of 30 data values were collected and summarised according to each section of
the rock slope outcrops (Figure 6).

722 Slope Stability 2020

 Assessment and implications for uncertainty in design

Figure 6 Manual measurement (Barton’s comb)

In CloudCompare, the discontinuities present in the rock mass was extracted using the FACET plugin. Of the
59 million points, 31 million points corresponded to the outcrop. Based on the dip/dip directions, a total of
3,852 facets were produced, and grouped after the FM approach in the plane segmentation process was
carried out. Two major discontinuity sets can be observed from Figure 6a, namely Set 1 (green) and Set 2
(blue). While the mean dip/dip direction for Set 1 was 046°/172°, it was 043°/189° for Set 2, as shown in
Figure 6b. For the entire rock slope, the mean dip/dip direction is 035°/186° as depicted in Figure 6c. The
stereograms shown in Figure 7 exported from CloudCompare, were plotted in dip vector mode. The values
extracted from this output have been verified using a scanline survey. Manual measurement of orientations
using scanline survey method was used to verify the accuracy of the extraction of geological planes by FACET
plugin in CloudCompare. The dip/dip direction of the same location was extracted from the software to
compare with that obtained manually. The data is tabulated in Table 4. There is a difference of 10° of dip
angle between the data extracted digitally from the CloudCompare tools and the manual measurement.

Figure 7 (a) Facets extracted from 3D dense cloud and Stereogram; (b) Green set; (c) Blue set

Slope Stability 2020 723

3D limit equilibrium method for rock slope stability analysis using generalised NS Kumar and MAM Ismail
anisotropic material model

Table 4 Comparison of dip/dip direction obtained manually and from software

Chainage Scanline survey Extraction from software

number
Dip angle (°) Dip direction (°) Dip angle (°) Dip direction (°)
2890-2900 70 230 68 223
80 100 82 95
40 160 44 164
80 80 71 89
2880-2890 85 260 85 253
50 160 46 162
90 80 82 85
2870-2880 70 80 79 91
40 190 39 179
50 170 44 180
2860-2870 50 170 45 175
90 250 85 243
60 180 61 183
2850-2860 50 170 52 175
75 40 83 45
65 222 66 232
80 70 82 85
2840-2850 40 230 36 220
30 210 34 213
60 80 49 77
40 220 35 205
2830-2840 50 70 49 77
40 180 43 185
2820-2830 45 150 44 154
30 170 35 167
2810-2820 50 160 48 164
40 160 43 168
2800-2810 80 30 77 39
60 150 54 159
30 180 43 190
70 240 66 232

724 Slope Stability 2020

 Assessment and implications for uncertainty in design

In situ observation and basic laboratory tests were conducted to obtain the rock input parameters. The phir
angle was obtained from the basic tilt test experiment. Table 5 shows the summary of input parameters used
for numerical modelling of the rock slope.

Table 5 Input parameter of rock material for anisotropic slope stability analysis

Material Parameters Comments

Unit weight 27 kN/m2
Uniaxial compressive
Rock mass 73.34 MPa
strength
(Generalised
Geological strength Calculated based on (Bieniawski 1989)
Hoek–Brown) 59
index (GSI) 1.5 JCond89 + RQD/2 = (1.5)16 + 70/2 = 59
Mi value 10 Spiritic Limestones
Disturbance factor (D) 0.8 Poor to Good Blasting
Joint compressive
52 MPa
strength
Weak joint JCS and JRC from correlated N value from
Joing roughness
(Barton–Bandis) 4 (Barton & Choubey 1977)
coefficient
Phir 33 degrees

5.2 2D and 3D slope stability analysis with and without anisotropy

Two cases of rock slope stability analysis were carried out: without anisotropic plane, which is a control set,
and with anisotropic plane. The rock slope stability for both cases was analysed by a deterministic method in
3D. Besides, cut-sections were extracted from 3D rock geometry respectively and analysed by deterministic
method in 2D. All the rock input parameters for both cases are the same except for the existence of
anisotropic plane. The mean dip/dip direction obtained was used as an input for the value of the anisotropic
plane where it causes a weakness in the strength of the rock slope. Figures 8a and 8b depicts the FoS of the
3D analysis with rock mass strength only. Figure 9 shows the surface failure of 2D rock geometry analysed
without anisotropic plane considering rock mass strength only. The least FoS of the entire rock slope is 5.446
(3D) and 4.716 (2D) respectively. This value indicates that the rock slope will be very stable if anisotropic
plane is not considered. Figure 10a, 10b and 11 illustrates the FoS of the 3D and 2D rock geometry analysed
with the anisotropic plane. The input dip/dip directions are 043°/189 and °046°/172°. The least deterministic
FoS of the entire rock slope obtained is 1.089 and 0.771 respectively. This value indicates that the rock slope
is less stable and tends to fail at any time.

Slope Stability 2020 725

3D limit equilibrium method for rock slope stability analysis using generalised NS Kumar and MAM Ismail
anisotropic material model

Figure 8 (a) Surface safety map; (b) Factor of Safety of rock slope without anisotropic plane in 3D

Figure 9 Factor of Safety of rock slope without anisotropic plane in 2D

726 Slope Stability 2020

 Assessment and implications for uncertainty in design

Figure 10 (a) Safety map of surface and (b) Factor of Safety of rock slope with anisotropic plane in 3D

Figure 11 Factor of Safety of rock slope with anisotropic plane in 2D

Slope Stability 2020 727

3D limit equilibrium method for rock slope stability analysis using generalised NS Kumar and MAM Ismail
anisotropic material model

The FoS of rock slope with and without the anisotropic plane in 3D and 2D analysis is tabulated in Table 6.
Besides, the 3D deterministic analysis gives higher FoS than that of 2D deterministic analysis. The difference
is because the 3D failure surface does not cross weak rock surface only, but also strong ones, neglecting the
over-conservative simplification assumed within the 2D sections. Furthermore, the lower FoS of rock slope
with anisotropic plane indicates that the orientations of the joints in the rock slope play an important role in
determining the strength of the rock mass. It will contribute to the weakness of the rock mass. Furthermore,
the lower FoS of rock slope with anisotropic plane indicates that the orientations of the joints in the rock
slope play an important role in determining the strength of the rock mass. The angle of orientation is
significant therefore it is a parameter that cannot be ignored when dealing with rock slope stability analysis.
The value of orientation is significant and therefore it is a parameter that cannot be ignored when dealing
with rock slope stability analysis. The rock slope orientation is also daylighting towards the valley which is
another factor that shows failure may occur. Based on the field observation, some port of the slope has failed,
this satisfies the results of with anisotropic plane. Planar sliding was observed as the most common failure
mode along the rock slope.

Table 6 Results of 2D and 3D slope stability analysis

Anisotropic plane Least deterministic Factor of FoS in 2D analysis (global

Safety (FoS) in 3D analysis minimum surface)
Without 5.446 4.716
With 1.089 0.771

6 Conclusion
SfM is one of the costs and time-effective methods in the aerial mapping of rock slopes. The 3D textured
geometric model of the rock slope produced was used for rock slope stability analysis together with the
orientations extracted act as the variables. The geological planes extracted from the high accuracy dense
point cloud using FACET plugin in CloudCompare is considered reliable and accurate. Two major discontinuity
sets and random discontinuities were identified during the extraction of the facets plane. Considering that
the anisotropic plane is important since the rock slope is heterogeneous, having various discontinuity
patterns, unpredictable and anisotropic, the geotechnical parameters of rock and discontinuities were
considered effective factors in the stability of rock slopes. Understanding the geology remains a tough
challenge as despite being detailed, all models may be wrong if the geology is not understood and
appropriately interpreted. Strength of anisotropy must be considered in the rock slope stability assessment
to portray the effects of discontinuities present in the rock mass. The utilisation of 3D rock slope stability
assessment was able to reduce the cost of slope remedial measures based on the failure spot. Application of
anisotropy model was able to demonstrate the weakness plane in the rock mass and the stability analysis
while neglecting the anisotropic condition may result in a wrong interpretation of the FoS.

Acknowledgement
The authors express their gratitude to Universiti Sains Malaysia fellowship program for providing financial
assistance in the form of scholarship. This work was also supported by the Universiti Sains Malaysia bridging
grant. The authors express their sincere appreciation to the lecturers, technicians and peers of the School of
Civil Engineering, Universiti Sains Malaysia, who have contributed towards this research.

References
Agam, MW, Hashim, MHM, Murad, MI & Zabidi, H 2016, ‘Slope sensitivity analysis using Spencer’s method in comparison with general
limit equilibrium method’, Procedia Chemistry, vol. 19, pp. 651–658, http://dx.doi.org/10.1016/j.proche.2016.03.066
Amadei, B 1996, ‘Importance of anisotropy when estimating and measuring in situ stresses in rock’, International Journal of Rock
Mechanics and Mining Sciences and Geomechanics Abstracts, vol. 33, issue 3, pp, 293–325.

728 Slope Stability 2020

 Assessment and implications for uncertainty in design

Asadi, M 2016, ‘Optimized Mamdani fuzzy models for predicting the strength of intact rocks and anisotropic rock masses’, Journal of
Rock Mechanics and Geotechnical Engineering, vol. 8, issue 2, pp. 218–224, http://dx.doi.org/10.1016/j.jrmge.2015.11.005
ASTM D4543 2001, Standard Practices for Preparing Rock Core Specimens and Determining Dimensional and Shape Tolerances, ASTM
International, https://doi.org/10.1520/D5079-08
ASTM D5873 1981, Determination of Rock Hardness by Rebound Hammer Method 1, ASTM International.
Bar, N & Weekes, G 2017, ‘Directional shear strength models in 2D and 3D limit equilibrium analyses to assess the stability of
anisotropic rock slopes in the Pilbara Region of Western Australia’, Journal and News of the Australian Geomechanics Society,
vol. 52, issue 4, pp. 91–104.
Barton, N & Choubey, V 1977, ‘The Shear Strength of Rock Joints in Theory and Practice’, Rock Mechanics Felsmechanik Mécanique
Des Roches, vol. 10, pp. 1–54, https://doi.org/10.1007/BF01261801
Bieniawski, ZT 1989, Engineering Rock Mass Classifications: A Complete Manual for Engineers and Geologists in Mining, Civil, and
Petroleum Engineering, Wiley, Hoboken.
Cala, M 2007, ‘Convex and concave slope stability analyses with numerical methods’, Archives of Mining Sciences, vol. 52, issue 1,
pp, 75–89.
Cheng, YM, Lansivaara, T & Siu, J 2008, ‘Impact of Convergence on Slope Stability Analysis and Design’, Computers and Geotechnics,
vol. 35, issue 1, pp, 105–113.
Cheng, YM, Liu, HT, Wei, WB & Au, SK 2005, ‘Location of Critical Three-Dimensional Non-Spherical Failure Surface by NURBS Functions
and Ellipsoid with Applications to Highway Slopes’, Computers and Geotechnics, vol. 32, issue 6, pp. 387–399.
Deere, DU & Miller, RP 1966, Engineering classification and index properties for intact rock, technical report, Air Force Weapons
Laboratory
Dewez, TJB, Girardeau-Montaut, D, Allanic, C & Rohmer, J 2016, ‘Facets : A Cloudcompare plugin to extract geological planes from
unstructured 3D point clouds’, International Archives of the Photogrammetry, Remote Sensing and Spatial Information
Sciences, pp. 799–804, https://doi.org/10.5194/isprsarchives-XLI-B5-799-2016
Francioni, M, Simone, M, Stead, D, Sciarra, N, Mataloni, G & Calamita, F 2019, ‘A new fast and low-cost photogrammetry method for
the engineering characterization of rock slopes’, Remote Sensing, vol. 11, issue 11, pp. 1–24.
Francioni, M, Salvini, R, Stead, D & Coggan, J 2018, ‘Improvements in the integration of remote sensing and rock slope modelling’,
Natural Hazards, vol. 90, issue 2, pp. 975–1004.
Gonzaga, GG, Leite, MH & Corthésy, R 2008, ‘Determination of anisotropic deformability parameters from a single standard rock
specimen’, International Journal of Rock Mechanics and Mining Sciences, vol. 45, issue 8, pp, 1420–1438.
Hoek, E, Carranza, C & Corkum, B 2002, ‘Hoek–Brown failure criterion – 2002 edition’, Proceedings of the 5th North American Rock
Mechanics Symposium and the 17th Tunnelling Association of Canada Conferenec, vol. 1, pp. 267–273.
Huang, CC & Tsai, CC 2000, ‘New Method for 3D and Asymmetrical Slope Stability Analysis’, Journal of Geotechnical and
Geoenvironmental Engineering, vol. 126, issue 10, pp. 917–927.
Ismael, MA, Imam, HF & El-Shayeb, Y 2014, ‘A Simplified Approach to Directly Consider Intact Rock Anisotropy,in Hoek-Brown Failure
Criterion’, Journal of Rock Mechanics and Geotechnical Engineering, vol. 6, issue 5, pp. 486–92,
http://dx.doi.org/10.1016/j.jrmge.2014.06.003
Kassa, HM & Steinar, N 2016, ‘Numerical models on anisotropy of rocks’, Proceedings of the 17th Nordic Geotechnical Meeting,
Icelandic Geotechnical Society, Reykjavik, pp. 587–596, https://www.ngm2016.com/uploads/2/1/7/9/21790806/064-057-
ngm_2016_-_numerical_models_on_anisotropy_of_rocks_kassa.pdf
Kim, D 2016, Study on the accuracy of rock surface roughness data using close range photogrammetry, PhD Thesis, Griffith University,
Brisbane.
Leong, EC & Rahardjo, H 2012, ‘Two and three-dimensional slope stability reanalyses of bukit batok slope’, Computers and
Geotechnics, vol. 42, pp. 81–88.
Mohamad, AB, Zain, AM & Bazin, NEN 2014, ‘Cuckoo search algorithm for optimization problems - A literature review and its
applications’, Applied Artificial Intelligence, vol. 28, no. 5, pp. 419–448, https://doi.org/10.1080/08839514.2014.904599
Nian, TK, Huang, RQ, Wan, SS & Chen, GQ 2012, ‘Three-dimensional strength-reduction finite element analysis of slopes: geometric
effects’, Canadian Geotechnical Journal, vol. 49, issue 5, pp. 574–588.
Nunes, ALLS 2002 ‘A new method for determination of transverse isotropic orientation and the associated elastic parameters for
intact rock’, International Journal of Rock Mechanics and Mining Sciences, vol. 39, issue 2, pp. 257–273.
Özvan, A, Dinçer, I, Acar, A & Özvan, B. 2014, ‘The effects of discontinuity surface roughness on the shear strength of weathered
granite joints’, Bulletin of Engineering Geology and the Environment, https://doi.org/10.1007/s10064-013-0560-x
Riquelme, AJ, Abellán, A, Tomás, R & Jaboyedoff, M 2014, ‘A new approach for semi-automatic rock mass joints recognition from 3D
point clouds’, Computers and Geosciences, vol. 68, pp. 38–52, https://doi.org/10.1016/j.cageo.2014.03.014
Salvini, R, Mastrorocco, G, Seddaiu, M, Rossi, D & Vanneschi, C 2017, ‘The use of an unmanned aerial vehicle for fracture mapping
within a marble Quarry (Carrara , Italy ): photogrammetry and discrete fracture network modelling’, Geomatics, Natural
Hazards and Risk, vol. 8, issue 1, pp, 34–52, http://dx.doi.org/10.1080/19475705.2016.1199053
Snow, DT 1969, ‘Anisotropic permeability of fractured media’, Water Resources Research, vol. 5, no. 6, pp, 1273–1289.
Westoby, MJ, Brasington, J, Glasser, NF, Hambrey, MH & Reynolds, JM 2012, ‘‘Structure-from-motion’ photogrammetry: a low-cost,
effective tool for geoscience applications’, Geomorphology, vol. 179, pp, 300–314, http://dx.doi.org/
10.1016/j.geomorph.2012.08.021
Wu, A 2012, Locating General Failure Surfaces in Slope Analysis via Cuckoo Search, Rocscience,
https://www.rocscience.com/help/slide2/pdf_files/theory/Cuckoo_Search.pdf

Slope Stability 2020 729

 3D limit equilibrium method for rock slope stability analysis using generalised NS Kumar and MAM Ismail
anisotropic material model

730 Slope Stability 2020

View publication stats