ADDIS ABABA UNIVERSITY

ADDIS ABABA INSTITUTE OF TECHNOLOGY

SCHOOL OF CIVIL AND ENVIRONMENTAL
ENGINEERING

RELIABILITY OF PROBABILISTIC FINITE ELEMENT

METHOD OVER DETERMINISTIC AND TRADITIONAL
PROBABILISTIC METHOD IN SLOPE STABILITY
ANALYSIS

A Thesis in Geotechnical Engineering

By
Eleyas Wolde Endashaw
April, 2022
Addis Ababa
 ADDIS ABABA UNIVERSITY
ADDIS ABABA INSTITUTE OF TECHNOLOGY
SCHOOL OF CIVIL AND ENVIRONMENTAL ENGINEERING

RELIABILITY OF PROBABILISTIC FINITE ELEMENT

METHOD OVER DETERMINISTIC AND TRADITIONAL
PROBABILISTIC METHOD IN SLOPE STABILITY
ANALYSIS

A Thesis in Geotechnical Engineering

By
Eleyas Wolde Endashaw

Advisor
Dr.-Ing. Tensay Gebremedhin

April, 2022
Addis Ababa

A Thesis
Submitted in Partial Fulfillment of the Requirements for the Degree of Master of Science
 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

ADDIS ABABA UNIVERSITY

ADDIS ABABA INSTITUTE OF TECHNOLOGY
SCHOOL OF CIVIL AND ENVIRONMENTAL ENGINEERING

RELIABILITY OF PROBABILISTIC FINITE ELEMENT METHOD

OVER DETERMINISTIC AND TRADITIONAL PROBABILISTIC

METHOD IN SLOPE STABILITY ANALYSIS

By: Eleyas Wolde Endashaw

The signatories have examined the thesis entitled ‘Reliability of Probabilistic Finite

Element Method over Deterministic and Traditional Probabilistic Method in Slope

Stability Analysis’ presented by Eleyas Wolde Endashaw, a candidate for the degree of

Master of Science and hereby certify that it is worthy of acceptance.

Advisor Signature Date

Internal Examiner Signature Date

External Examiner Signature Date

Chair person Signature Date

i
Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

ii
 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

UNDERTAKING

I attest that research work titled “Reliability of Probabilistic Finite Element Method over
Deterministic and Traditional Probabilistic Method of Slope Stability Analysis” is my
work. The work has not been presented elsewhere for assessment. Where material has been
used from other sources it has been properly acknowledged / referred.

………………………..

Eleyas Wolde Endashaw

iii
 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

ABSTRACT

Slope Stability is one of the crucial topics in geotechnical engineering to be investigated

well because oftentimes slope exists naturally or formed artificially during construction of
various civil engineering structures which needs to be at equilibrium. For that matter, there
are various methods of analysis. For the last couple of decades, it is observed hard for
Engineers to use Probabilistic Slope Stability Analysis due to many reasons. But these
days, technological advancements which led to having more rigorous software packages
made it be possible to use in the analysis and its merit. The aim of this thesis is to show
the use of more reliable method of slope stability analysis for hypothetical back slopes in
Addis Ababa where red clay soil is dominant and that is achieved by comparing methods
based on a factor of safety, probability of failure, and reliability index.

Limit Equilibrium Method (LEM) and Numerical Method are used for this purpose. Both
deterministic and probabilistic analyses are carried on by modeling seven slope geometries
with different input parameters from Addis Ababa ring road III project and hypothetical
slopes using a two-dimensional Slope Stability analysis software called SlideV6.0. A
Global Minimum type of analysis are carried on by using the Slide software. For finite
element analysis and probabilistic finite element analysis, RS2 software from Rocscience
is used. Uncertainty is accounted for in a better way in probabilistic analysis by defining
parameters as a random variable in link with other features.

It is shown by this study that, Probabilistic Finite Element Analysis (PFEA) seeks out the
most critical slip surface than that of the finite element method. In addition, even if the
factor of safety found from the deterministic analysis is greater than one still a probability
of failure is observed. Moreover, it is attested that the factor of safety is not the only
measure for slope stability, but other measures exist while using probabilistic analysis.

Key words: Slope, Probabilistic analysis, FEM, PFEA, Slope Stability

iv
 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

ACKNOWLEDGMENTS

Above all, all Glory and Honor to the Almighty and then, I take the opportunity to
acknowledge persons and institutions who have contributed a lot to my research. My
sincere gratitude goes to my advisor Dr.-Ing Tensay Gebremedhin for his guidance with
full of patience, comments, and his invaluable role in collecting all necessary data and
equipment. And then, I am extremely grateful to KAAD and Alumni members in Bon and
Addis Ababa. It was once not viable to make it without their scholarship offer.

I would like to express thanks also to the Ethiopian Roads Authority Research Center
(ERARC) for allowing me to use the raw data they have collected for this research and am
thankful to the Ethiopian Railway Center of Excellence and all staff members for letting
me use the computer lab of the center. Moreover, my heartfelt gratitude to my professors
and all staff under the chair of Geotechnical Engineering at Addis Ababa University.

Finally, I highly appreciate my beloved family, ministry of Science and Higher Education,
Community of St. John Addis Ababa, and all good friends who have contributed much to
this research in different ways.

v
 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

TABLE OF CONTENETS

UNDERTAKING ........................................................................................................... III

ABSTRACT.................................................................................................................... IV

ACKNOWLEDGMENTS ............................................................................................... V

LIST OF TABLES ......................................................................................................... IX

LIST OF FIGURES ......................................................................................................... X

LIST ABBREVIATIONS ............................................................................................. XII

CHAPTER 1 INTRODUCTION ............................................................................. 13

1.1 Background ....................................................................................................... 13

1.2 Statement of the Problem.................................................................................. 14
1.3 Objective of the Study ...................................................................................... 15
1.3.1 General objective ........................................................................................ 15

1.3.2 Specific objectives ...................................................................................... 15

1.4 Scope of the Study ............................................................................................ 16

1.5 Significance of the study .................................................................................. 16
1.6 Organization of the Thesis ................................................................................ 16
CHAPTER 2 LITERATURE REVIEW ................................................................. 18

2.1 Definition and Background............................................................................... 18

2.2 Shear Strength for Slope Stability .................................................................... 19
2.3 Slope Stability Analysis.................................................................................... 20
2.3.1 Limit Equilibrium Analysis ........................................................................ 21

2.3.1.1 Swedish Slip Surface Method ............................................................. 23

2.3.1.2 Ordinary Method of Slices .................................................................. 24

2.3.1.3 Simplified Bishop’s Method ............................................................... 24

2.3.1.4 Morgenstern – Price Method ............................................................... 25

2.3.1.5 Spencer’s Method ................................................................................ 26

2.3.1.6 Janbu’s Method ................................................................................... 27

vi
 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

2.3.1.7 Sarma’s Method .................................................................................. 28

2.3.1.8 Probabilistic Slope Stability Analysis ................................................. 29

2.3.1.8.1 First Order Second Moment (FOSM) Method ................................ 32

2.3.1.8.2 Point Estimate Method .................................................................... 32

2.3.1.8.3 First Order Reliability Method (FORM) ........................................ 33

2.3.1.8.4 Monte Carlo simulation Method ..................................................... 33

2.3.2 Numerical Method ...................................................................................... 35

2.3.2.1 Finite element Method......................................................................... 35

2.3.2.2 Probabilistic Finite Element Method ................................................... 36

2.4 Summary of Related Studies ............................................................................ 37

CHAPTER 3 MATERIALS AND METHODS ...................................................... 42

3.1 Introduction....................................................................................................... 42
3.2 Study Area ........................................................................................................ 43
3.3 Regional Geology ............................................................................................. 43
3.4 Seismicity of the region .................................................................................... 45
3.5 Topography ....................................................................................................... 45
3.6 Climate.............................................................................................................. 46
3.9 Data Collection ................................................................................................. 48
3.10 Software used for Analysis ............................................................................... 48
3.11 Analysis Methods ............................................................................................. 51
3.12 Input parameters ............................................................................................... 52
3.13 Modeling ........................................................................................................... 53
3.13.1 Slide V6.0 ................................................................................................... 53

3.13.2 RS2 V11.0................................................................................................... 59

CHAPTER 4 RESULTS AND DISCUSSIONS ..................................................... 64

4.1 Limit Equilibrium Analysis .............................................................................. 64

4.2 Finite Element Analysis .................................................................................... 67
4.2.1 FEM by using RS2...................................................................................... 67

4.2.2 Probabilistic FEM Analysis by Using RS2 ................................................ 69

vii
 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

CHAPTER 5 CONCLUSIONS AND RECOMMENDATIONS .......................... 72

5.1 Conclusions....................................................................................................... 72
5.2 Recommendation .............................................................................................. 72
REFERENCES ............................................................................................................... 73

APPENDIX A .................................................................................................................. 80

APPENDIX B .................................................................................................................. 98

APPENDIX C ................................................................................................................ 112

APPENDIX D ................................................................................................................ 126

viii
 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

LIST OF TABLES

Table 2.1 Values of Coefficient of Variation (V) for Geotechnical Properties and In Situ
Tests (Duncan et al., 2000) ............................................................................................... 38
Table 2.2 Approximate guidelines for design soil property variability ............................ 38
Table 2.3 Average values μ and coefficient of variation COV for the active angle of
internal friction (Alamanis, 2017) .................................................................................... 39
Table 2.4 Average values of μ and coefficient of variation COV for active cohesion
(Alamanis, 2017) .............................................................................................................. 39
Table 2.5 Average values μ and coefficient of variation COV for the unit weight
(Alamanis, 2017) .............................................................................................................. 40

Table 3.1 Bed rock acceleration ratio (ES EN 1998-1:2015 Design of Structures for
earthquake resisistance , 2015) ......................................................................................... 45
Table 3. 2 Average Temperature (Climate- Ethiopia, n.d.) .............................................. 47
Table 3. 3 Average Precipitation (Climate- Ethiopia, n.d.) .............................................. 47
Table 3.4 Defined COV from different kinds of literature ............................................... 58

Table 4.1 Factor of safeties, global minimum analysis type ............................................ 65

Table 4. 3 Summary of factor of safety result from RS2.................................................. 68
Table 4.4 Summary of findings ........................................................................................ 71

ix
 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

LIST OF FIGURES

Figure 2.1 Causes of slope failure (Civil Seek, 2018) ...................................................... 19

Figure 2.2 Stability of slopes by Culmann method (Murthy, 2002) ................................. 22
Figure 2.3 Common types of slope failures (Civil Seek, 2018) ....................................... 23
Figure 2.4 Forces on ordinary method slice (Aryal, 2006)............................................... 24
Figure 2.5 Forces on Simplified Bishop’s slice (Aryal, 2006) ......................................... 25
Figure 2.6 Forces on Morgenstern-Price slice (Aryal, 2006) ........................................... 26
Figure 2.7 Forces on Spencer's slice (Aryal, 2006) .......................................................... 27
Figure 2.8 Forces on Janbu's simplified slice (Aryal, 2006) ............................................ 27
Figure 2.9 Forces on Janbu's generalized slice (Aryal, 2006) .......................................... 28
Figure 2.10 Slope geometry illustrating Janbu’s direct method (Aryal, 2006) ................ 28
Figure 2.11 Very erratic spatial structure (upper right) and a highly continuous structure
(lower right), both with similar histograms (El-Ramly, Morgenstern, & Cruden, 2002). 31

Figure 3.1 Location of landslide affected areas in the highlands of Ethiopia, Modified after
the works of the various researchers (Woldearegay, 2013) .............................................. 42
Figure 3.2 Site Location (Google Earth) .......................................................................... 43
Figure 3.3 Geological map of Addis Ababa City (Geology of Ethiopia, 1998) ............... 44
Figure 3.4 Random Variable Samples used in Probabilistic Analysis (Rocscience web
help, 2021) ........................................................................................................................ 50
Figure 3.5 Soil geometry Addisu gebeya 1, slide V6.0 .................................................... 54
Figure 3.6 Soil geometry Addisu gebeya 2, slide V6.0 .................................................... 54
Figure 3.7 Soil geometry Kolfe 2, slide V6.0 ................................................................... 55
Figure 3.8 Convergence plot Addisu gebeya 2 ................................................................. 56
Figure 3.9 Automatic Calculation of Hu coefficient (Rocscience web help, 2021) ......... 57
Figure 3.10 Discretized and meshed model by RS2 Asko ............................................... 60
Figure 3. 11 Discretized and meshed model by RS2 Kolfe 1 ........................................... 60
Figure 3.12 Quadratic Triangular element........................................................................ 61

Figure 4.1 Global minimum analysis showing global minimum slip surface for Bishop,
Addisu gebeya 2 ............................................................................................................... 65

x
 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Figure 4.2 Global minimum analysis showing all slip surfaces for Bishop, Addisu gebeya
2 ........................................................................................................................................ 65
Figure 4.3 Critical SRF showing contour of the maximum shear strain, Kolfe 1 ............ 68
Figure 4.4 Deformed shape and Boundaries, Addisu gebeya 2 ........................................ 69
Figure 4.5 Maximum total displacements at critical SRF, Addisu gebeya 2 ................... 69
Figure 4.6 Mean critical SRF, showing the contour of Maximum shear plastic strain .... 70
Figure 4.7 Deformed shape and boundaries, Addisu gebeya 2 ........................................ 70

xi
 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

LIST ABBREVIATIONS

AASHTO – American Association of State Highway and Transportation Officials

COV – Coefficient of Variation

EGS - Ethiopian Geological Survey

ERA – Ethiopian Road Authority

FEM – Finite Element Method

FOSM – First Order Second Moment

FS – Factor of Safety

GM – Global Minimum

LAS – Local Average Subdivision

LEM - Limit Equilibrium Method

MCS – Monte-Carlo Simulation

PFEA – Probabilistic Finite Element Analysis

PLC – Private Limited Company

PSSA – Probabilistic Slope Stability Analysis

RHS – Right Hand Side

RI- Reliability Index

RS – Rock and Soil

RS2 – A 2D Finite Element Software from Rocscience

SS – Sliding Surface

SSRF - Shear Strength Reduction Factor

USGS - United States Geological Survey

xii
 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

CHAPTER 1 INTRODUCTION

1.1 Background

Slope stability analysis is one of the main areas of interest to geotechnical engineers in
analyzing and designing different geotechnical structures like retaining structures,
tunnels, shoring systems, reinforced earth structures, and so on. Analysis of the safety
level of slopes is very important for proper geotechnical risk management and
assessment. There are classical or conventional and advanced ways of slope stability
analysis used for both natural and man-made slope stability analysis. But the
conventional methods made many assumptions and have limitations in simulating the
real soil conditions. However, the advanced methods of analyses are found to be more
realistic in considering the variability of the soil properties to assess the reliability and
risk associated with the projects since the soil is an inherently heterogeneous, non-single-
phase material that has natural variability and uncertainty in its properties.

To simulate the uncertainty of soil properties in recent years, probabilistic approaches

are more and more widely used for the safety assessment of slopes as mentioned by El-
Ramly et.al., (2002) and are more reliable in providing an acceptable design criterion.
The degree of safety can be determined through probabilistic slope stability analyses with
the consideration of uncertainties and variability in material properties, as well as in
environmental factors, like the fluctuation of the groundwater table due to seasonal
variation. It is noted by various researchers, Oguz et.al., (2017), that a slope with a
deterministic safety factor (FS) that we gate from the conventional methods larger than
1.00 may have a probability of failure greater than 0%.

Moreover, an advanced and rigorous probabilistic approach for modeling the influence
of statistically described soil properties on design outcomes in geotechnical engineering
have emerged during the early ‘90s (Fenton & Vanmarcke, 1990) and (Griffiths &
Fenton, 1993). Among these, finite element based probabilistic analysis and Random
Finite Element Method (RFEM) can be mentioned. The latter one encompasses both
random field theory and the finite element method. The method was applied to several
areas of geotechnical engineering including slope stability analysis according to Griffiths
et.al., (2000), (2004). The Local Average Subdivision method (LAS) proposed by Fenton

MSc Thesis Page 13

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

and Vanmarcke (Fenton G. A., 1990) is used for generating the random fields. According
to Allahverdizadeh et.al., (2015), in traditional probabilistic analyses, spatial variability
is neglected by implicitly assuming perfect correlation, which does not necessarily result
in a conservative estimate of the probability of failure.

This research is aimed to show the reliability of Probabilistic Finite Element Method
(PFEM) in slope stability analysis over deterministic and other traditional (conventional)
probabilistic analysis by using various slopes in Ethiopia in link with various
stakeholders like Ethiopian Roads Authority (ERA) and Geological Survey of Ethiopia
(GSE).

Hence, to achieve the stated goal, data is gathered for the selected specific site and
appropriate software programs are chosen. In order to investigate the instability, both
Limit Equilibrium Method (LEM) and Finite Element Method (FEM) are used. For the
limit equilibrium method (deterministic and probabilistic analyses), Slide software from
Rocscience is used and for finite element and probabilistic finite element method of
analysis, RS2 software is used also from Rocscience.

From the analysis followed, it is found that the probabilistic finite element method seeks
out the most critical slip surface relative to other methods of slope stability analysis. In
addition, even in LEM, still probabilistic analysis gives a lesser safety factor than that of
deterministic analysis when a probabilistic analysis is computed in overall slope type.
And also, from the deterministic analysis, a higher factor of safety that seems safe is
observed, but not, since the probability of failure exceeded zero.

1.2 Statement of the Problem

Conventional limit equilibrium methods are under use for the assessment of slope
stability in routine practice in parts of the world as well as in our country Ethiopia and
the soil profiles are often assumed to be uniform and homogenous which is not true in
reality. In addition to this, the conventional slope stability analyses are often performed
with in deterministic analysis framework were single estimates or characteristic values
for soil parameters are used.

MSc Thesis Page 14

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

To consider the spatial variability and uncertainty of soil properties most of the time a
higher value of factor of safety is adopted in deterministic methods as stated by Chok
et.al., (2015). As a result, the conventional approach may not be reliable to estimate the
potential failure of slope.

But with all these limitations, we are using the conventional and deterministic analysis
approach in our country. As it was mentioned in previous section, we adopt a higher
factor of safety which is uneconomical to avoid risk, and as it has already been researched
by many for instance Oguz et.al., (2017) even if a high factor of safety is adopted, there
may be a probability of failure. In many parts of our country, we see many man-made
and natural slope failures and some of them are disastrous for life and infrastructures.
Method of analysis can be one of the ways to minimize risk and failure.

1.3 Objective of the Study

1.3.1 General objective

The general objective of this research is to investigate the reliability of static Probabilistic
Finite Element Method (PFEM) of analysis over that of deterministic LEM and
traditional probabilistic LEM in slope stability analysis in terms of Factor of Safety (FS),
Probability of failure (Pf), Reliability index (RI) and Critical failure surfaces in different
hypothetical slopes in Ethiopia.

1.3.2 Specific objectives

• Investigate the effect of defining material properties as a random variable in

probabilistic slope stability analyses.

• Investigate the effect of analysis type on safety factor values and stability condition
of slopes.

• Investigate the effect of assuming a slip surface in advance on the value of safety
factor.

MSc Thesis Page 15

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

1.4 Scope of the Study

Slope stability analysis is a vast area of research in geotechnical engineering. In this

research, only a few LEM of analyses are considered and from the probabilistic analysis
method, the Monte Carlo and Latin hypercube are used.

Besides, the study is limited to Static slope stability analysis, in other words, the dynamic
slope stability analysis is not included in the scope.

1.5 Significance of the study

The Ethiopian Geological Survey (GSE) and Ethiopian Roads Authority (ERA) are
reporting many slope failures in a different part of the country. For instance, around
Dessie, Wondo Genet, Blue Nile Gorge, Jimma basin, Goffa, Sawla, Jinka and Wolaita
area are few among many. As a country, we are investing much money for roads and
other infrastructures but due to slope stability problems their functionality does not last
long. A well-proven analysis method in addition to other aspects like understanding the
possible causes of failure, has a significant contribution to minimize failures which also
impact the economy positively.

1.6 Organization of the Thesis

This thesis is organized in 5 chapters and appendices, each presenting condensed

information entitled with. A summary is given for each below.

Chapter 1: This chapter presents a general introduction to the thesis, from what the
research idea is initiated, it states the aim of the thesis, all the ways how the research is
done, its scope, and limitations.

Chapter 2: In this chapter, numerous pieces of literature are reviewed about the topic
including related works done by former scholars. The key points in slope stability
analysis and the different types of slope stability analyses are discussed in this chapter
and especially probabilistic analysis.

MSc Thesis Page 16

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Chapter 3: Depicts materials and methods utilized in this research. It presents the
general background of the study area and its various formations including
characterization of the land slide. It states how data is gathered and describes briefly the
background of software programs used. This chapter gives the general procedure and
process of how the analysis is carried on step by step for each methodology mentioned
in chapter one.

Chapter 4: The results from the analysis carried on are interpreted briefly in this chapter.
Outputs are summarized and discussed by using plots from the software used which
elaborate the findings. Comparison between analyses is also made in this chapter.

Chapter 5: This is the last chapter containing major findings and recommendations. The
conclusions made from the values obtained from LEM of analysis and the Numerical
method of analysis are indicated. Finally, data logs, output data, material boundary,
model geometry coordinates, and all necessary information used in the software are
presented from Appendix A-D.

MSc Thesis Page 17

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

CHAPTER 2 LITERATURE REVIEW

2.1 Definition and Background

A slope is defined as a surface of which one end or side is at a higher level than another;
a rising or falling surface. It is an inclined soil mass without support (Salunkhe, Bartakke,
Chvan, & Kothavale, 2017). The slope of the earth classified as Natural, the one that is
formed by natural factors and exists in nature and Manmade that is formed by human
actions which include the sides of cuttings, the slopes of embankments constructed for
roads, railway lines, canals, etc. and the slopes of earth dams constructed for storing
water (Murthy, 2002).

According to Salunkhe et al., (2017), these natural and manmade slopes may be grouped
under finite and infinite slopes depending on the extent of inclination. when a slope has
a limited length of inclination it can group under finite slope and when a slope represents
the boundary surface of a semi-infinite soil mass or ideally the inclination is extended to
unlimited length and the soil properties for all identical depths below the surface are
constant the slope is termed as an infinite slope. The slope is used to canal banks, road
cuts, landfills open-pit mining, excavations, the earthen dam, railway formation, highway
embankment, levees, etc.

Something involving a downward and outward movement of the whole mass of soil that
involves in the failure we call it a slide. gravitational forces and force due to seepage
water, erosion, sudden lowering of water near to a slope, external loading, geological
feature, earthquake forces are the main triggering forces for slope failure within the soil
(Figure 2.1). They may also fail due to adjacent excavation or undercutting of its foot, or
due to gradual disintegration of the structure of the soil due to weathering (Salunkhe,
Bartakke, Chvan, & Kothavale, 2017).

The movement of the slope is either by expansion and contraction of soil due to the
variation in temperature called soil-creep or sudden movement of large soil mass refers
to mass slides or flow slides which occur when the soil starts to flow outward and
downward from its natural position (Teferra A. and Leikun M., 1999).

MSc Thesis Page 18

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Figure 2.1 Causes of slope failure (Civil Seek, 2018)

One of the oldest, controversial, and arguably the most researched and the least
understood topic of geotechnical engineering is the stability of slopes. In their book titled
Soil Strength and Slope Stability Duncan et al., (2014) said that “Evaluating the stability
of slopes in the soil is an important, interesting, and challenging aspect of civil
engineering”. Among various works in this field in terms of different approaches
(theoretical, experimental, analytical, statistical, and numerical), Terzaghi’s (1950) work
entitled “Mechanism of Landslides” is well known. Researchers are looking for an
advanced approach today since the solution of slope stability problems requires the
understanding of analytical methods and their application, the awareness of known
methods and their drawbacks becomes a crucial topic.

2.2 Shear Strength for Slope Stability

Effective or total stress parameters can be used to assess the stability of a slope. For
cohesionless soils forming the slope material, the effective stress analysis is best suitable
since drainage occurs quickly after loading (or unloading) of the slope disregarding
exceptional circumstances. For the case of clay-type soils, with low pore pressures,
MSc Thesis Page 19
 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

permeability and hence there is time-dependent stability of the slope (ETİZ, 2019) an
effective stress approach with drained shear strength parameters can be applicable if the
pore-water pressures can be estimated fairly close. If it is not possible to estimate or
determine the pore water pressure, the total stress type of analysis may have to be
employed. In practice, it is common to use the total stress analysis for short-term stability
problems and the effective stress principle is used to assess the long-term stability,
assuming that all excess pore pressure generated during loading will be fully dissipated
(Abramson, Thomas S. Lee, Sharma, & Glenn, 2001).

2.3 Slope Stability Analysis

The reason for slope stability analysis is to assess the safe design of manmade or natural
slopes and the equilibrium conditions. The slope is the resistance of inclined surface at
some angle to failure by sliding or collapsing which may cause loss of life and asset. It
is, therefore essential to check the stability of proposed slopes.

The stability of a slope can be checked in terms of factor of safety; for analysis either
factor of safety for shearing strength i.e., the ratio of shear strength (shearing resistance)
to the shear stress (mobilized shear strength) along sliding surface or factor of safety for
cohesion i.e., the ratio of available cohesion intercept and the mobilized cohesion
intercept might be used (Murthy, 2002). And also, there is a factor of safety for friction
which is the ratio of available frictional strength to mobilized frictional strength.

During the analysis of slope, there should be a distinction between finite and infinite
slopes (in this paper only analysis of finite slope is reviewed). Thanks to the advancement
of modern methods of soil testing and stability analysis, a safe and economical design of
slope are admissible (Salunkhe, Bartakke, Chvan, & Kothavale, 2017).

Slope stability analysis is broadly classified as:

A. Static Slope Stability Analysis

B. Dynamic Slope Stability Analysis

MSc Thesis Page 20

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

A. Static Analysis
i.Limit equilibrium method:
❖ Swedish slip circle method of analysis
❖ The ordinary method of slices
❖ Simplified bishop’s method of analysis
❖ Janbu’s method of analysis
❖ Spencer’s method of analysis
❖ Sarma method of analysis
❖ Taylor’s stability number
❖ Probabilistic analysis based on LEM
ii.Numerical method:
❖ Finite difference method
❖ Finite Discrete method
❖ Finite Element method
❖ The probabilistic FE method
iii.Numerical method of modeling:
❖ Continuum modeling
❖ Discontinuum modeling
❖ Hybrid/coupled modeling

2.3.1 Limit Equilibrium Analysis

The limit equilibrium method is a conventional slope stability assessment approach,

which considers force or moment equilibria of the soil mass remaining above a probably
assumed failure surface (ETİZ, 2019). The basic assumption of the limit equilibrium
method is that Coulomb's failure criterion is satisfied. Culmann (1875) assumed the
failure surface as a plane shown in Figure 2.2 for the analysis of slopes which is mainly
of interest because it serves as a test of the validity of the assumption of plane failure
along the assumed failure surface. It starts from the known or assumed values of forces
acting on the free body diagram taken from the slope and then the shear resistance is
calculated, this calculated value is compared with the available shear strength (Murthy,
2002).

MSc Thesis Page 21

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Figure 2.2 Stability of slopes by Culmann method (Murthy, 2002)

There is a circular failure surface as the investigation shows. When soil slips along a
circular surface, such a slide may be termed as a rotational slide. Such failures can be
classified as slope failure, toe failure, and base failure as shown in Figure 2.3.

MSc Thesis Page 22

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Figure 2.3 Common types of slope failures (Civil Seek, 2018)

In the LEM method, the soil mass above the slip surface is assumed to be rigid at the
time of failure, the available shear strength is assumed to be mobilized at the same rate
at all points of the slip surface. This implies that FS is constant throughout the failure
surface. In LEM factor of safety can be calculated in the limit force and moment
equilibriums (ETİZ, 2019). Some of the LEMs are briefly discussed in the next section.

2.3.1.1 Swedish Slip Surface Method

It is also known as the circular arc method or φ=0 method of analysis (Teferra A. and
Leikun M., 1999). In 1922 the Geotechnical Commission was appointed by the Swedish
State Railways to investigate solutions following a costly slope failure. The analysis
followed for this project has laid a base for the development of the limit equilibrium
analysis method. The method has become known as the Swedish Slip Circle Method. In
this method, a cylindrical shape that looks a circle in cross-section is assumed.

MSc Thesis Page 23

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

This method assumes zero resultant interslice force, a circular slipping surface, and the
equilibrium of overall moment is satisfied but the equilibrium of individual slice moment,
equilibrium of both vertical and horizontal forces is not satisfied.

2.3.1.2 Ordinary Method of Slices

It is known as the “Ordinary Method of Slices” or Fellenius ‘method developed by

Fellenius in 1936. In any method of slices, the soil mass above the failure surface is
subdivided into vertical slices, and the stability is calculated for each slice. Fellenius’
Method simplifies the equation by assuming that the forces acting on the sides of each
slice cancel each other (Fellenius, 1936).

Figure 2.4 Forces on ordinary method slice (Aryal, 2006)

Where, S = shear force acting at the base of a slice (kN), N’= effective base normal force
acting on SS (kN)

The ordinary method of slices doesn’t satisfy the force equilibrium of both the entire
slide mass and individual slices and doesn’t satisfy the equilibrium of moment of
individual slice but overall moment equilibrium. It assumes a circular slipping surface.

2.3.1.3 Simplified Bishop’s Method

This method accounts for the interslice normal forces, which increase the accuracy of
calculating the factor of safety (FS). The method of slices, therefore developed is known
as the “Simplified Bishop’s Method” (Bishop A.W., 1955). However, Bishop’s Method
still cannot satisfy all the conditions of static equilibrium for example equilibrium of
horizontal forces, equilibrium of moment for an individual slice. All interslice forces are

MSc Thesis Page 24

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

assumed to be zero, which reduces the number of unknowns by (n-1). Remaining (4n-1)
unknowns, overdetermined solution as horizontal force equilibrium not satisfied for a
slice (ETİZ, 2019).

Figure 2.5 Forces on Simplified Bishop’s slice (Aryal, 2006)

Where, S = shear force acting at the base of a slice (kN), N’= effective base normal force
acting on SS (kN), W weight of each slice or total sliding mass (kN), E = interslice normal
force (kN).

2.3.1.4 Morgenstern – Price Method

It is known as the “Morgenstern-Price Method”. Morgenstern and Price developed this

alternative method for analyzing slides with a non-circular failure path (Morgenstern &
Price, 1965). This method satisfies all equilibrium of forces and moments. unlike the
method of analysis discussed earlier, it is possible to assume any shape of the sliding
surface. The morgenstern-price method sets an arbitrary mathematical function to
describe the direction of the interslice forces.
𝑥
𝜆𝑓(𝑥) = (2.1)
𝐸
Where λ is a constant scale factor of the assumed function to be evaluated when solving
for the FS and ƒ(x) = interslice force function that varies continuously along the slip
surface.

MSc Thesis Page 25

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Figure 2.6 Forces on Morgenstern-Price slice (Aryal, 2006)

In general, Morgenstern-Price Method considers both interslice forces, assumes an

interslice force function, f(x), allows selection for interslice force function, computes FS
for both force and moment equilibrium (Aryal, 2006).

2.3.1.5 Spencer’s Method

This method is a complete equilibrium method known as Spencer’s Method, which

satisfies the equilibrium of both force and moment. As a result, the FS calculated by this
method should be more precise. In spencer’s method, it is possible to assume a non-
circular slip surface which is useful because many slope failures do not have exactly
circular failure surfaces (Spencer, 1967).

MSc Thesis Page 26

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Figure 2.7 Forces on Spencer's slice (Aryal, 2006)

2.3.1.6 Janbu’s Method

Janbu has developed three different methods of slope stability analysis namely, “Janbu’s
Simplified Method” (Janbu, 1954a), Janbu’s generalized method, and Janbu’s direct
method. a non-circular failure surface is possible in Janbu’s simplified method which
considers the interslice forces with an assumption of interslice/ tangential forces are equal
to zero, but includes a correction factor to compensate for the interslice forces.

Figure 2.8 Forces on Janbu's simplified slice (Aryal, 2006)

However, in Janbu’s generalized method both interslice forces and assume a line of thrust
to determine a relationship for interslice forces. satisfies both force and moment

MSc Thesis Page 27

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

equilibriums, handles complex geometry and failure surfaces, is an advanced method

among LE methods.

Figure 2.9 Forces on Janbu's generalized slice (Aryal, 2006)

Finally, Janbu’s direct method is based on a series of charts as a powerful tool to carry
out slope stability analysis. Besides, surcharge load, groundwater, and tension crack
loading conditions are possibly integrated during analysis (Aryal, 2006).

Figure 2.10 Slope geometry illustrating Janbu’s direct method (Aryal, 2006)

2.3.1.7 Sarma’s Method

In 1973, Sarma came up with a method for general blocks or a non‐vertical slice.
considers both interslice normal and shear forces, satisfies both moment and force
equilibrium, and relates the interslice forces by a quasi‐shear strength equation.

𝑇 = 𝑐ℎ + 𝐸𝑡𝑎𝑛𝜑 (2.2)

MSc Thesis Page 28

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Where, c, 𝜑 are shear strength parameters, and ‘h’ is height.

Overall, limit equilibrium method:

➢ gives different results in different methods because all methods make different
assumptions.
➢ uses an upper bound solution
➢ the sliding block is assumed to be rigid
➢ distribution of FS is taken as constant throughout the failure zone
➢ distribution of stress is taken as constant throughout the failure zone
➢ it doesn’t give the displacement of the sliding block

2.3.1.8 Probabilistic Slope Stability Analysis

The variability of soil property and uncertainty is not considered in conventional

deterministic slope stability analysis explicitly but relies on conservative parameters and
designs to deal with uncertain conditions. However, Probabilistic analyses allow
uncertainty to be quantified and incorporated rationally into the design process. Besides,
it can be applied in LEM and FEM of analysis. Probabilistic slope stability analysis
(PSSA) was first introduced into slope engineering in the 1970s (El-Ramly, Morgenstern,
& Cruden, 2002).

According to El-Ramly et al., (2002), there are four reasons for practicing engineers
facing difficulties in using PSSA:

❖ less comfortability dealing with probabilities due to limited understanding of

statistics and probability.
❖ due to the common misconception that probabilistic analyses require significantly
more data, time, and effort than deterministic analyses.
❖ Due to misconception that PSSA replaces deterministic but rather they are
supplementary to each other.
❖ limited studies regarding benefits and illustration of PSSA and
❖ acceptable probabilities of unsatisfactory performance (or failure probability) are
ill-defined, and the absence of a link between a probabilistic assessment and a
conventional deterministic assessment.

MSc Thesis Page 29

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

In the technical sense, variability can be defined as an observable manifestation of

heterogeneity of one or more physical parameters and/or processes. It is an inherent
property to a natural material. The uncertainty reflects the decision (or necessity) to
recognize and address the observed variability in one or more soil properties of interest
(Phoon & Kulhawy, 1999). It arises from the difficulty in measuring properties.

Geotechnical uncertainties may be aleatory uncertainty (inherent variability) or epistemic

uncertainty (measurement error, statistical estimation error, and model error). The main
source of uncertainty in soil is from spatial variability of soil properties. Soil is variable
in composition and properties from place to place even in a homogenous soil layer. This
variability is due to factors like composition, deposition condition, stress history, and
physical and mechanical decomposition process. Figure 2.11 compares the spatial
variability of two artificial sets of data. The upper plot is highly erratic and the data are
almost uncorrelated, whereas the lower plot is characterized by high spatial continuity.
The pattern of soil variability is characterized by the autocorrelation distance or scale of
fluctuation. A large autocorrelation length implies soil homogeneity and a short
autocorrelation length shows erratic variability.

MSc Thesis Page 30

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Figure 2.11 Very erratic spatial structure (upper right) and a highly continuous structure (lower
right), both with similar histograms (El-Ramly, Morgenstern, & Cruden, 2002)

The other measure of uncertainty is in spatial averaging. Rather than soil properties at
discrete locations, the performance of a structure is often controlled by the average soil
properties within a zone of influence. Slope failure is more likely to occur when the
average shear strength along the failure surface is insufficient rather than due to the
presence of some local weak pockets. The uncertainty of the average shear strength along
the slip surface, not the point strength, is therefore a more accurate measure of
uncertainty (El-Ramly, Morgenstern, & Cruden, 2002).

As stated by Griffiths et al., (2015) slope stability analysis has received more attention
in the case of probabilistic geotechnical analysis, it seems likely that a very significant
bibliography is now available like Matsuo and Kuroda (1974), Alonso (1976), Tang et
al. (1976) and Vanmarcke (1977).

According to El-Ramly et al., (2005) probabilistic procedures for slope stability analysis
vary in assumptions, limitations, capability to handle complex problems, and
mathematical complexity and they can be grouped in two categories: approximate

MSc Thesis Page 31

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

methods (the FOSM method, the point estimate method) and Monte Carlo simulation.
However, now a days another sampling method called Latin Hypercube is emerged.
Approximate methods make assumptions that often limit their application to specific
classes of problems and do not provide any information about the shape of the probability
density function, so the failure probability can only be obtained by assuming a parametric
probability distribution of the factor of safety (typically normal or log-normal). Estimates
of low probabilities, required for safe structures, are sensitive to the assumed distribution.
Some of the probabilistic slope stability Analysis methods are discussed briefly in the
fore coming sections.

2.3.1.8.1 First Order Second Moment (FOSM) Method

It is called FOSM because it uses the first order of the Taylor approximation terms about
the second moments (mean and variance). This method is an approximate approach that
uses the first order approximation of the mean, standard deviation, and variance of the
performance function (the distribution of the factor of safety) which depends on the first
order Taylor’s Series expansion and neglect the higher order terms (AKBAŞ, 2015). In
slope stability works, this function is the factor of safety which is calculated by the
method of slices. FOSM calculates the reliability index (β) from the mean value ( μ𝐹𝑆 )
and standard deviation (σ𝐹𝑂𝑆 ) of a factor of safety.

𝜇𝐹𝑆 − 1
𝛽= (2.3)
𝜎𝐹𝑆
The advantage of this method is that it can show the relative contributions of each random
variable to the performance function. This can be useful since the designers can focus on
the most affecting variable. This is not provided in many of the other methods. On the
other hand, the demerit of the method is to deal with the partial derivative of the
performance function, especially if it has a complicated form (AKBAŞ, 2015).

2.3.1.8.2 Point Estimate Method

It is known as Rosenblueth’s method. It’s an approximate and easy to use and

straightforward method where the designers are not required to know much of a
probability theory. The method approximates the moments of the performance function
that contains random variables. As it is first and second moments of a function are mean

MSc Thesis Page 32

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

and variance, respectively and the square root of the variance is the standard deviation.
initially, it is seen as an over approximate approach. However, its satisfactory accuracy
is shown by Baecher & Christian (2005) in several numerical case studies. The
correlation between variables is possible in addition to its simplicity. But this method is
better for problems with few variables to save computation time (AKBAŞ, 2015).

2.3.1.8.3 First Order Reliability Method (FORM)

In FOSM and point estimate method, first, for the estimation of performance function,
only the mean and variance of the random variables and their linear combination are
used, and second, in knowing the form of distribution of performance function the
probability of failure of performance function can be obtained from reliability index
information. To overcome these two assumptions FORM developed.

The first order reliability method (FORM) is a process that can be used to determine the
probability of a failure given the distribution data and limit state function. The method is
based on the Hasofer-Lind reliability index (βHL) (Hasofer and Lind 1964), which can be
described as the distance, in standard deviation units, between the most probable set of
values and the most probable set of values that causes a failure. Calculation of this value
is an iterative process, finding the minimum value of a matrix calculation subject to the
constraint that the values result in a system failure. The reliability index can be
determined and then the probability of failure, Pf, (Griffiths, Fenton, & Denavit, 2007).
the distance in standard deviation units between the most probable set of random
variables (the means), and the most probable set of random variables that causes a failure
is called the reliability index, and an iterative process is used for its determination.

𝑥𝑖 − 𝜇𝑖 𝑇 𝑥𝑖 − 𝜇𝑖 𝑇
𝛽𝐻𝐿 = 𝑚𝑖𝑛 √( ) + [𝑅]−1 ( ) (2.4)
𝑔→0 𝜎𝑖 𝜎𝑖

𝑥𝑖 −𝜇𝑖 𝑇
Where, ( ) is the vector of random variable values reduced to standard normal space,
𝜎𝑖

[𝑅] is the correlation matrix of the values of the variables.

2.3.1.8.4 Monte Carlo simulation Method

The Monte Carlo method is a powerful slope stability analysis developed in 1949 by John
von Neumann and Stanislaw Ulam ( (Eckhardt, 1987), (Fishman, 1995)). Monte Carlo

MSc Thesis Page 33

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Simulation (MCS) is based on a random sampling procedure (sometimes called pseudo-

random numbers (Zhang, Ji, & Xu, 2020) in which input variables affecting the
performance function (distribution of FS) is randomly selected from a region of interest
and results on the performance function is evaluated. Monte Carlo simulation is a
powerful tool for slope stability risk analysis.

In this method, an iterative process using deterministic methods of slope stability analysis
is applied. Monte Carlo simulation is a popular method of slope stability risk analysis
much used by engineers because of its simplicity and no need for comprehensive
mathematical and statistical knowledge (Sharma, 2016). The first step of a Monte Carlo
simulation is to identify a deterministic model where multiple input variables are used to
estimate a single value outcome. Step two requires that all variables or parameters be
identified. Next, the probability distribution for each independent variable is established
for the simulation model, (i.e., normal, beta, lognormal, etc.). Next, a random trial
process is initiated to establish a probability distribution function for the deterministic
situation being modeled.

Generalized steps can be defined as follows: (AKBAŞ, 2015)

A. Fixing geometry of the slope to be analyzed and the most likely (mean) values of
the required soil parameters.
B. Choosing the shear strength soil model (Mohr-Coulomb, Hoek-Brown, etc.)
C. Determination of the randomly treated input parameters (i.e., unit weight,
cohesion, internal friction angle, pore water pressure coefficient, etc.)
D. For each random variable, choosing the relevant COV from the literature unless
information on the variability of site-specific soil is sufficient.
E. Choosing the statistical distribution (normal, lognormal, etc.) of the random
parameter.
F. Choosing several required analyses (N) and sampling methods.
G. Choosing the shape of a slip surface (circular, non-circular, etc.) if necessary.
H. Carrying the slope stability analysis N-times.

This procedure will result in a statistically distributed N-times factor of safeties,

probability of failure, critical probabilistic failure surface, and reliability index (AKBAŞ,
2015).

MSc Thesis Page 34

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

The limitation of this method is that generated number may not cover the necessary
region of interest depending on the values parameters defined, seed, and the number of
realizations. In addition to that, depending on the probability of failure of the
performance function, necessary random number realization can be too much.

2.3.2 Numerical Method

2.3.2.1 Finite element Method

In the finite element method (FEM) the limitations of the limit equilibrium method are
solved. In FEM, the soil continuum is divided into discrete units called “finite elements”.
The finite elements are interconnected at nodal points and the prescribed boundaries of
the mass continuum. In typical geotechnical applications, the displacement method of
formulation of the FEM is utilized to calculate displacements, stresses, and strains at the
nodal points.

In LEM statics of force and moment equilibrium are considered but in FEM stress-strain
relationship is used like constitutive law. Stress redistributions can be computed since it
is based on the stress-strain relationship. Besides, compatibility between structural
members and soil media can be done without much problem due to its meshing process.
The principal uses of the finite element method as stated by Duncan (1996) for design
are:

➢ it provides estimates of displacements and construction pore water pressures.

➢ it provides a displacement pattern that may show potential and possibly complex
failure mechanisms. Once a potential failure mechanism is recognized, the factor
of safety against a shear failure developing by that mode can be computed using
conventional limit equilibrium procedures.
➢ it provides estimates of mobilized stresses and forces.
➢ the finite element method may be particularly useful in judging what strengths
should be used when materials have very dissimilar stress-strain and strength
properties, i.e., where strain compatibility is an issue.
➢ the FEM can help identify local regions where “overstress” may occur and cause
cracking in brittle and strain softening materials. An essential input to the stability

MSc Thesis Page 35

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

analyses for reinforced slopes is the force in the reinforcement. The FEM can
provide useful guidance for establishing the force that will be used.

The shear strength reduction factor (SSRF) method is employed in FEM to analyze slope
which provides a very quick and reasonable estimate of stability. This method is based
on the reduction of cohesion (c) and friction angle (tanφ) of the soil.

𝐶 tanφ
SSRF = 𝐹𝑆 = = (2.5)
𝐶𝑟𝑒𝑑𝑢𝑐𝑒𝑑 tan𝜑𝑟𝑒𝑑𝑢𝑐𝑒𝑑

The parameters are reduced in steps until the soil mass fails. Different computer
programs which are based on the finite element method, use the stress reduction method
in the analysis of slope. One of the main advantages of the Shear Strength Reduction
Technique (SSRF) is that the safety factor emerges naturally from the analysis without
assuming any particular form of failure mechanism in advance.

The shear strength reduction factor ‘F’ increases incrementally until the global failure of
the slope reaches, implies, under a physically real convergence criterion, finite element
calculation diverges. The lowest factor of safety of slope lies between the shear strength
reduction factor F at which the iteration limit is reached and the immediately previous
one. The procedure described at this moment can predict the factor of safety within one
loop and can be easily implemented in a computing code (Nakamura, 2008). However,
FEM is time consuming and complex because no assumptions are made to simplify the
problem and further processing is required because FEM does not provide a value for the
overall factor of safety of the computed stresses.

2.3.2.2 Probabilistic Finite Element Method

Probabilistic Finite Element Method (PFEM) determine the effect of uncertainty or

variability of input parameters on the results of the finite element analysis by defining
input parameters as random variables. The PFEM combines elastic-plastic finite element
analysis with random field theory in slope stability analysis (Allahverdizadeh, ASCE,
Griffiths, Fenton, & ASCE, 2015).

The PFEM is used in link with Monte Carlo simulations with many different realizations
of the soil properties onto a finite element mesh in which the stability analysis is repeated

MSc Thesis Page 36

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

until the probabilities relating to output quantities of interest become statistically

reproducible. In the case of slope stability analysis, the probability of failure is defined
by dividing the number of realizations in which the slope failed by the total number of
realizations as stated by Griffiths et al., (2007). Because of avoiding of assumption of
shape and location of failure surface PFEM is more realistic and powerful probabilistic
slope stability analysis.

The unique feature of this method is the ability to seek out the weakest and most critical
path and it doesn’t stick with a certain shape of failure surface while accounting for
spatial variability of soil (Griffiths & Fenton, 2004). It incorporates the main advantage
of FEM in this regard.

2.4 Summary of Related Studies

Oguz et al., (2017), investigated the effects of the Coefficient of Variation (COV) and
the Cross-Correlation of Shear Strength Parameters. COV is defined as the ratio between
standard deviation and mean. In their study, they considered uncertainty of soil properties
by different levels of coefficient of variation. They used the limit equilibrium method for
slope stability analyses with a normal statistical distribution of geotechnical material
properties. The results showed that the Probability of failure (Pf) and the critical failure
surface are significantly influenced by the COV level, cross correlation of shear strength
parameters, and the traditional FS level of the slopes. Moreover, the inverse relation
between FS and Pf is demonstrated to be nonlinear and the COV level has a significant
effect on this relationship.

It is shown that large values of COV, i.e., significant variation in material properties, can
cause large values of Pf for slopes with FS greater than unity and Pf decreases with
increasing COV for FS less than unity, and deterministic FS alone will not be sufficient
to indicate this danger in slope stability. As cohesion increases the friction angle
decreases, this is termed as a negative cross correlation which is considered in their study.

There are suggested values of COV by different researchers in the literatures. But care
should be given in using these values since it is not clearly presented how these values
are recorded. Duncan and others provided a summary of COV values for various property

MSc Thesis Page 37

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

or in situ test results, only properties related to this study are presented in Table 2.1
(Duncan, Member, & ASCE, 2000).

Table 2.1 Values of Coefficient of Variation (V) for Geotechnical Properties and In Situ
Tests (Duncan et al., 2000)
Property or in situ test result COV Reference
(%)
Unit weight (𝛾) 3–7% Harr (1984), Kulhawy (1992)

Buoyant unit we (𝛾𝑏 ) 0–10% Lacasse and Nadim (1997),

Duncan (2000)
Effective stress friction angle (φ′) 2–13% Harr (1984), Kulhawy (1992)

Undrained shear strength (𝑆𝑢 ) 13–40% Harr (1984), Kulhawy (1992),

Lacasse and Nadim (1997),
Duncan (2000)

Phoon and Kulhawy conducted the effect of COV on the analysis of slope stability and
they reported a summary of values for different soil types and different soil properties
which can be used as a guideline during limitation on the availability of sufficient data
(Phoon & Kulhway, 1999).

Table 2.2 Approximate guidelines for design soil property variability

(Phoon & Kulhway, 1999)
Design Test Soil type Point Spatial
property COV(%) avarage (%)
𝑆𝑢 (UC) Direct (lab) Clay 20–55 10–40
𝑆𝑢 (UU) Direct (lab) Clay 10–35 7–25
𝑆𝑢 (CIUC) Direct (lab) Clay 20–45 10–30
𝑆𝑢 (field) VST Clay 15–50 15–50
φ’ Direct (lab) Clay, sand 7–20 6–20
φ’ (TC) 𝑞𝑇 Sand 10–15 10

Where,
𝑆𝑢 , undrained shear strength; 𝑆𝑢 (field), corrected 𝑆𝑢 from vane shear test(VST); φ’,
effective stress friction angle; φ’cv, constant-volume φ; TC, triaxial compression; UC,
unconfined compression test.

MSc Thesis Page 38

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Nikolaos Alamanis in recent years summarized the value of mean and COV for different
types of soil properties. Here only some of the soil properties are presented, for further
inquiry, the reader can refer (Alamanis, 2017).

Table 2.3 Average values μ and coefficient of variation COV for the active angle of
internal friction (Alamanis, 2017)
Researcher Year Mean μ COV
Harr 1987 2% - 13%
Kalhawy 1992 2% - 13%
Phoon and his colleagues 1995 20 - 40 (deg) 5% - 15%
Lacasse and his colleagues 1997 2% - 5%
Suchomel 2010 21 (deg) 8%
Phoon and his colleagues 1999 21-40 (deg) 5% - 15%
Duncan 2000 2% - 13%
Jeremic and his colleagues 2007 2% - 5%
Griffiths and his colleagues 2002 35 (deg) 5% - 50%
El Ramley and his colleagues 2002 35 (deg) 5.60%
Schweiger 2005 35 (deg) 0

Table 2.4 Average values of μ and coefficient of variation COV for active cohesion
(Alamanis, 2017)
Researcher Year Mean μ COV
Griffiths and his colleagues 2002 24KN/m2 30%
Suchomel 2010 10KN/ m2 21%
Harr 1987 20%
Cherubini 1997 20%-30%
Li and his colleagues 1987 40%

MSc Thesis Page 39

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Table 2.5 Average values μ and coefficient of variation COV for the unit weight
(Alamanis, 2017)
Researcher Year Mean μ(kN/m3) COV
Harr 1987 1%-10%
Phoon and his colleagues 1995 13-20 <10%
Smith and his colleagues 2004 20 0%
Duncan 2000 14-20 <10%
Wang and his colleagues 2010 20 6%
Hicks and his colleagues 2002 20 0%
Griffiths and his colleagues 2002 20 0%
Schweiger 2005 20 0%

El-Ramly et al., (2002), have investigated the impact of uncertainty on the reliability of
slope design and performance assessment. According to their study, conventional slope
practice based on the factor of safety cannot explicitly address uncertainty, thus
compromising the adequacy of projections. Probabilistic techniques are rational means
to quantify and incorporate uncertainty into slope analysis and design. They have used
Microsoft® Excel 97 and @Risk as a tool based monte carlo simulation. An important
conclusion of thier study is that probabilistic analyses can be applied in practice without
an extensive effort beyond that needed in a conventional analysis. They believe that
combining conventional deterministic slope analysis and probabilistic analysis will be
beneficial to slope engineering practice and will enhance the decisionmaking process.

Nejan Huvaj and Emir Ahmet Oğuz (2018), investigated a well-documented landslide
case study to demonstrate the importance of probabilistic approach in slope stability.
They investigated the effects of considering variability in material properties and
compared deterministic and probabilistic slope stability analyses results. Deterministic
limit equilibrium, probabilistic limit equilibrium, and probabilistic finite element
analyses are conducted for Lodalen landslide in Oslo, Norway and they compared results
with each other. They have investigated factor of safety, the probability of failure and the
most critical failure surface with and without statistical cross-correlation of soil’s shear
strength parameters. They comeup with the conclusion that the critical failure surface in
a slope can be different indeterministic LEM, probabilistic LEM, and probabilistic FEM

MSc Thesis Page 40

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

analyses. It would be best to identify the most critical failure surface by making use of
all available methods, including probabilistic approaches.

R.K Sharma (2016), have worked on probabbilistic analysis and reach to a conclusion
that traditional slope stability analysis is limited to the use of single valued parameters to
analyze a slope’s characteristics. Thus, traditional analysis methods yield single valued
estimates for factor of safety of a slope’s stability. Yet, the inherent variability of the soil
characteristics which affect slope stability shows that the stability of a slope is a
probabilistic rather than a deterministic situation. In other words, the stability of a slope
is a random process which is dependent on the relative distribution of controlling soil
parameters. For a natural slope, the stability deciding parameters vary considerably
throughout the extent of slope. In his study, the variability of soil properties and their
effect on stability of a natural slope studied by incorporating the probabilistic analysis
using Monte Carlo simulation and deterministic analysis using Geo-Studio and PLAXIS.
The results obtained from probabilistic approach can be used to determine the probability
of failure corresponding to a particular of factor of safety and an allowable risk criterion
can be used to establish a consistent target for the design process.

MSc Thesis Page 41

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

CHAPTER 3 MATERIALS AND METHODS

3.1 Introduction

The Ethiopian Geological Survey and Ethiopian Roads Authority have made some effort
even though there is no comprehensive record of landslides and their consequence is
made in detail in Ethiopia. Despite all challenges many scholars have reported on slope
instability problems in different parts of the country like Dessie, Abay gorge, Jimma
basin, Goffa, Gilgel Gibe-II, and Sodo-Shone areas, shale hill slopes, Tarmaber and
surrounding area, wollo, wondogenet, and Tekeze hydro power project as shown in
Figure 3.1.

Figure 3.1 Location of landslide affected areas in the highlands of Ethiopia, Modified
after the works of the various researchers (Woldearegay, 2013)

MSc Thesis Page 42

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

3.2 Study Area

The chosen study areas are found in the capital city of Ethiopia, Addis Ababa where red
clay soil is dominantly exists. A typical red clay locations were selected among the sub
cities of Addis Ababa namely, Kolfe Keranio, Arada, and Gulele sub cities. The specific
locations were Addisu Gebeya, Kolfe, Asko and Arada.

Figure 3.2 Site Location (Google Earth)

3.3 Regional Geology

Addis Ababa is positioned on the western edge of the main Ethiopian Rift Valley.
Normal faults down thrust towards the rift characterize the Mio-Pliocene boundary,
which shows Mio-Pliocene volcanic. The massive fault marks the upper (outside) edge
of the margin immediately north of the Addis Ababa-Ambo Road, it runs roughly east-
west. The lower limit runs northeast to southwest, parallel to the major fissure systems
in the area from Nazareth to Awash Station, the rift-floor (Akalu, 2017).

MSc Thesis Page 43

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Figure 3.3 Geological map of Addis Ababa City (Geology of Ethiopia, 1998)

Trachy-basalts (around general wingate school), Ignimbrites, and Basalts are the
principal occurrences with in western areas of Addis Ababa, as per geological map of
Addis Ababa. Addis Ababa basalt, primarily alkaline and olivene basalt, and Addis
Ababa ignimbrite cover parts of Gulele, Addis Ketema, and Arada. Wechecha's core
volcanics, Entoto's pyroclastic trachytic lava flows cover the majority of Kolfe
Keraniyo's subcity (Akalu, 2017).

As proposed by many researchers, Miocene-Pleistocene volcanic sequence from

bottommost to topmost is Addis Ababa Basalts, Bofa Basalts, Alaji Basalts, Entoto
Silicics, and Nazareth Group.

Addis Ababa basalt is porphyritic in texture which dominantly cover the central part the
city. Bafo basalts found in the southeastern region of Addis Ababa characterized by big
vesicles that are filled by calcite. It has a restricted location outcrop from southwards of
Akaki river basin. Alaji basalts has various textures from porphyritic to aphanitic covers
the north and northeastern part of the city. Entoto Silicic found at the bottom of entoto
hills composed of rhyolites and trachytes. The Nazareth group is composed of three units

MSc Thesis Page 44

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

identified as lower welded tuff, aphanitic basalt and upper welded tuff which covers
south of Filwoha fault and extended towards Nazareth (Haile Sellasie, 1989).

3.4 Seismicity of the region

According to the seismic hazard map of Ethiopia, parts of Addis Ababa city like bole and
kaliti town lies in zone three. Our country Ethiopia adopted an earthquake code of design
standards for buildings based on the seismic zone map by Pierre Gouin. Consequently,
Ethiopia is broadly subdivided into four seismic zones depending upon more than 500
historic earthquake records covering the year between 1400-1974. The seismic hazard
within each zone has been assumed to be constant through local variations are expected.

The seismic hazard map of Ethiopia as provided in the Ethiopian Building Code Standard
is based on a 100-year return period or approximately 50% of being exceeded in 50 years.
Each seismic zone 1 to 4 is assigned a constant bedrock acceleration ratio α0 of 0.03,
0.05, 0.07, or 0.1, respectively, however, Zone 0 is considered seismic free. In this map,
the study site belongs to Zone 1 with α0 = 0.03.

Based on the new code ES EN 1998-1:2015 the seismic hazard map is divided into five
zones based on bedrock acceleration with a reference return period of 475 years with a
10% probability of exceedance in 50 years. The study area was found in the same zone
according to the new code.

Table 3.1 Bed rock acceleration ratio (ES EN 1998-1:2015 Design of Structures for
earthquake resisistance , 2015)

Zone 5 4 3 2 1 0

α0 = ag /g 0.2 0.15 0.1 0.07 0.04 0

3.5 Topography

When we come to the overall topography of Addis Ababa, the city is surrounded by
volcano formed mountains like Wechecha mountain at the west, Furi mountain at the
southwest, Entoto mountain at the north, and Yerer mountain at southeast side. Light

MSc Thesis Page 45

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

brown and yellowish-brown soils are common in the north, northeast and northwest part
due leaching occurred in steep slope. However, soils in the southern, southeastern, central
and western parts are dark grey soils in color.

3.6 Climate

In Ethiopia, the climate varies mostly with altitude, and it goes from the hot and arid
climate of the lowlands to the cool climate of the plateau. In general, there are five
climatic zones, "Kur" (Alpine), above 3000 meters above sea level; "Dega" (Temperate),
2300 meters to around 3000 meters; "Weina Dega" (Subtropical), 1500 meters to around
2300 meters; "Kolla" (Tropical), 800 meters to about 1500 meters; and "Bereha"
(Desert), less than 800 meters. The majority of Addis Ababa is classified as Weina Dega
(subtropical) (Akalu, 2017).

3.7 Temperature

Addis Ababa, the capital, is located at 2,300 meters (7,500 feet) above sea level, with
altitudes ranging from 2,100 to 2,700 meters (7,000 to 9,000 feet) in different sections of
the city, and enjoys a warm temperature. From November to February, when lows drop
below 10 °C (50 °F), nights are cool, even cold, while days are pleasant, with highs
around 23/25 °C (73/77 °F), except in July, August, and September, when highs drop to
approximately 20/21 °C (68/70 °F) during the rainy season. As is customary in Ethiopia,
the months of March and May are the warmest of the year, albeit by a few degrees.

MSc Thesis Page 46

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Table 3. 2 Average Temperature (Climate- Ethiopia, n.d.)

Addis Ababa - Average temperatures (1991-2020)
Min Max Mean Min Max Mean
Month
(°C) (°C) (°C) (°F) (°F) (°F)
January 9 24 16.3 48 75 61.4
February 10 25 17.5 50 77 63.5
March 12 25 18.6 54 77 65.5
April 13 25 18.8 55 77 65.8
May 13 25 18.6 55 76 65.5
June 12 23 17.5 54 73 63.4
July 12 21 16.4 54 69 61.6
August 12 20 16.2 54 69 61.2
September 12 21 16.5 53 71 61.7
October 10 23 16.6 51 73 61.8
November 9 23 16.1 49 73 61
December 8 23 15.5 47 73 59.9
Year 11 23.1 17 51.9 73.5 62.5

3.8 Rainfall

Rainfall averages 1,160 mm (46 in) per year, with a peak from June to September, which
is the only particularly rainy period. From November to February, there is minimal rain
and only a few showers; from March to May, afternoon showers become more common,
occurring 7/10 days per month, and practically every day in July and August.

Table 3. 3 Average Precipitation (Climate- Ethiopia, n.d.)

Addis Ababa - Average precipitation
Month Millimeters Inches Days
January 13 0.5 3
February 30 1.2 5
March 60 2.4 7
April 80 3.1 10
May 85 3.3 10
June 140 5.5 20
July 280 11 27
August 290 11.4 26
September 150 5.9 18
October 25 1 4
November 7 0.3 1
December 7 0.3 1
Year 1165 45.9 132
MSc Thesis Page 47
 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

3.9 Data Collection

The data used for this research have been taken from two previous research works carried
on in Addis Ababa red clay soil. Among these researches the first one was aimed to
investigate the effect of remolding on mechanical behavior of Addis Ababa red clay soil
(Lukas, 2010) and the second one was aimed to investigate the engineering
characteristics of red clay soil in western Addis Ababa (Akalu, 2017). The former study
was carried on sites located at Addisu Gebeya and Kolfe and two samples were used for
each, the shear strength parameters determined from triaxial test (CU) and for the later
research, samples were taken from Asko and Arada area, direct shear test was used to
determine the shear strength parameters. Some soil parameters which are unable to get
from these researches, literature recommended values are used with caution by
considering the soil profile, each layer material description and soil log.

3.10 Software used for Analysis

By using the advantage of computer technology advancement nowadays software

programs are widely used in the advanced Engineering world. There are different types
of geotechnical software programs used to analyze the stability of any natural and
manmade slopes. Slide from Rocscience is one of the various software packages used to
analyze slope stability problems and used for this research. It is a 2D slope stability
analysis program for assessing the safety factor or likelihood(probability) of failure. It is
possible to assume both circular and non-circular failure surfaces in soil or rock slopes.
Outstandingly, Slide is so user friendlily to utilize, and yet complex models can be
created and analyzed rapidly and effectively. Modeling of external loading, groundwater,
and support can be done in different ways in the Slide (Rocscience web help, 2021).

Slide analyzes the stability of slip surfaces utilizing vertical slice limit equilibrium
methods (e.g., Bishop, Janbu, Spencer, etc.). Every single slip surface can be analyzed
or search methods can be connected to find the critical slip surface for a given slope. It
is possible in Slide to carry out deterministic (safety factor) or probabilistic (probability
of failure) investigations.

MSc Thesis Page 48

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Program modules in the Slide are three namely, model, compute and interpret module.
filling and editing the model boundaries, loads, material properties, groundwater
conditions, slip surface definition, and saving the input file is done in the model module
and it is a pre-processing program. The second one is a compute engine or program
module that runs from the Slide Model program. Finally, data visualization and
interpretation of the Slide analysis results is performed by Slide interpret program as
post-processing. This includes slope stability results as well as finite element
groundwater results if you have performed a groundwater seepage analysis (Rocscience
web help, 2021).

One of the unique features of this software is that there is a way to carry out a probabilistic
analysis and this is a good insight to know that factor of safety is not the only criteria for
failure as explained in the literature review section. In a conventional slope stability
analysis, it is expected that the values of all model input parameters are precisely known.
For a given slip surface, a single value of the safety factor is calculated. This sort of
investigation can be termed a deterministic analysis. In reality, the values of many input
parameters for a given slope instability are not very well-identified, that’s why the
probabilistic method to the analysis of slope stability can be suitable for better risk
minimization.

In a probabilistic slope stability analysis, assigning statistical distributions to model input

parameters, like material properties, support properties, loads, water table location, etc.
are possible by the user. In doing so the degree of uncertainty in the value of the
parameters is accounted for in a better way. Based on the user-defined statistical
distributions and by entering the appropriate parameters for the distribution (standard
deviation, minimum and maximum values, correlation coefficient) input data samples
are randomly generated. A given slip surface may then have numerous diverse values of
safety factors calculated. This allows the distribution of safety factors, from which a
probability of failure for the slope can be calculated (Rocscience web help, 2021).

Based on the selected statistical distribution, sampling method, and several samples(N)
defined N values of random variables are generated. By loading a new set of random
variable samples, and re-running the analysis Probabilistic Analysis is carried out that is

MSc Thesis Page 49

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

repeated N times as shown in Figure 3.4 where N is the number of Samples defined in a
project setting.

One iteration of probabilistic analysis

Figure 3.4 Random Variable Samples used in Probabilistic Analysis (Rocscience web help, 2021)

There are two sampling methods in Slide namely Monte Carlo and Latin Hypercube
sampling which determines statistical inpoout distributions for the random variables. The
Monte Carlo sampling technique uses accidental numbers to sample from the input data
probability distributions. It is applied to a wide variety of problems involving random
behavior, in geotechnical engineering (Rocscience web help, 2021). Latin Hypercube
sampling method is based upon "stratified" sampling with random selection within each
stratum. This results in a smoother sampling of the probability distributions.

In probabilistic slope stability analysis critical probabilistic slip surface can be shown
which will not be always the same as the critical deterministic slip surface (i.e., the slip
surface with the lowest safety factor when all input parameters are equal to their mean
value). In general, the critical probabilistic surface and the critical deterministic Surface
(i.e., the deterministic global minimum slip surface) can be different surfaces.

MSc Thesis Page 50

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

In addition to the factor of safety, probability of failure and reliability index are used as
a measure of safety in probabilistic analysis. By definition probability of failure is the
ratio of the number of failed surfaces to a total number of slip surface in the analysis and
the Reliability Index characterizes the number of standard deviations which separate the
mean Factor of Safety from the critical Factor of Safety (= 1) (Rocscience web help,
2021).

The second software used for this research is RS2 from Rocscience. RS2 is a powerful
2D finite element package used for various engineering projects like excavation design,
slope stability, groundwater seepage, probabilistic analysis, consolidation, and dynamic
analysis in soil and rock applications (Rocscience, 2021).

RS2 has a similar graphical user interface with that of Slide except for some and the
major feature which makes RS2 different from the Slide is that it is finite element-based
software. Unlike that of Slide, there is no predetermined failure surface in RS2 but rather
it works under the principle of strength reduction factor. It is possible to use different soil
models like Mohr-Coulomb and Generalized Hoek-Brown and other constitutive models.
Besides, it is also possible to work on probabilistic analysis on finite element mesh which
integrates the advantage of defining parameters as random variables than deterministic
ones.

3.11 Analysis Methods

By using those two software packages discussed in brief, two types of analysis are carried
out. These are deterministic and probabilistic analyses as summarized below.

❖ Slide V6.0 from Rocscience – Limit equilibrium method

✓ Deterministic analysis
✓ Probabilistic analysis
❖ RS2 V11.0 – Constitutive equations
✓ Finite element analysis
✓ Probabilistic finite element analysis

MSc Thesis Page 51

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

3.12 Input parameters

The following parameters are defined for both software programs to be used wherever
necessary. These data are taken from previous research works on the same location as
mentioned the research works in data collection section. But, Modulus of elasticity,
poison’s ratio, and porosity values are taken from different kind of literature (Bowles,
1997) and summarized useful parameters of the university of Stanford (Stanford
University, 2014) cause during soil investigation, these parameters are not considered.

➢ Kolfe 1
Cohesion (C’) = 23.4 kN/m2
Internal friction angle (ϕ’) = 16.7⁰
Unit weight (γsat) = 18 kN/m3
Elastic modulus = 100 000 kPa
Poison’s ratio = 0.4
Porosity = 0.5
➢ Kolfe 2
Cohesion (C’) = 19.2 kN/m2
Internal friction angle (ϕ’) = 23⁰
Unit weight (γsat) = 18 kN/m3
Elastic modulus = 50 000 kPa
Poison’s ratio = 0.4
Porosity = 0.5
➢ Kolfe 3
Cohesion (C) = 102.1 kN/m2
Internal friction angle (ϕ’) = 0⁰
Unit weight (γsat) = 22 kN/m3
Elastic modulus = 100 000 kPa
Poison’s ratio = 0.4
Porosity = 0.5
➢ Addisu gebeya 1
Cohesion (C’) = 20 kN/m2
Internal friction angle (ϕ’) = 15.5⁰
Unit weight (γsat) = 20kN/m3

MSc Thesis Page 52

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Elastic modulus = 50 000 kPa

Poison’s ratio = 0.4
Porosity = 0.5
➢ Addisu gebeya 2
Cohesion (C’) = 15.3 kN/m2
Internal friction angle (ϕ’) = 21.3⁰
Unit weight (γsat) = 19.5 kN/m3
Elastic modulus = 50000 kPa
Poison’s ratio = 0.4
Porosity = 0.5
➢ Arada
Cohesion (C’) = 71kN/m2
Internal friction angle (ϕ’) = 6⁰
Unit weight (γsat) = 20 kN/m3
Elastic modulus = 50 000 kPa
Poison’s ratio = 0.4
Porosity = 0.5
➢ Asko
Cohesion (C’) = 19 kN/m2
Internal friction angle (ϕ’) = 23⁰
Unit weight (γsat) = 19 kN/m3
Elastic modulus = 50 000 kPa
Poison’s ratio = 0.4
Porosity = 0.5

3.13 Modeling

3.13.1 Slide V6.0

The hypothetical soil geometry is modeled. It is a homogenous soil stratum. Ground

water table is assumed by considering the ground water depth variation in Addis Ababa
as shown in the Slide V6.0 model in Figure 3.5-3.7. A total of six geometries were
modeled with varying dimensions in the horizontal direction and vertical direction. For

MSc Thesis Page 53

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

the assignment of external loading, the worst combination of truck design load is taken
including the design lane load and its position is defined as per AASHTO design
requirement.

Figure 3.5 Soil geometr y Addisu gebeya 1, slide V6.0

Figure 3.6 Soil geometry Addisu gebeya 2, slide V6.0

MSc Thesis Page 54

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Figure 3.7 Soil geometry Kolfe 2, slide V6.0

Both deterministic and probabilistic analyses are carried out by using the advantage of
Slide to run both at the same time. While working on the model important project settings
are fixed. Among the various methods of analysis Bishop’s Simplified,
GLE/Morgenstern-Price, Janbu simplified and Spencer methods are selected by
considering the satisfaction of equilibrium of forces and moments. The groundwater
analysis is set to water surface implies that water surface is used to define the pore
pressure conditions for each soil type.

For the probabilistic analysis, the sampling method is set to Latin hypercube which uses
random numbers to sample from the input data probability distributions with 1000
samples which converges as shown in the Figure 3.8 and sensitivity is checked. The
method is based upon "stratified" sampling with random selection within each stratum.
The analysis method is global minimum type of analysis.

MSc Thesis Page 55

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Figure 3.8 Convergence plot Addisu gebeya 2

The failure surface is assumed to be circular with an auto refine critical search method
by which different slip surfaces are generated each time when analysis runs.

There are different strength type models available in Slide for modeling the shear strength
of the various materials. For this research Mohr-Coulomb, sterngth type is selected
(Equation 3.1).

𝜏 = 𝑐′ +  n tanφ′ (3.1)

As mentioned erlier, the groundwater method is set as water surface for pore pressure
calculation by using Equation 3.2.

𝑢 = 𝛾𝑤 ℎ𝐻𝑢 (3.2)

Where, 𝑢 = pore pressure, 𝛾𝑤 = unit weight of water, ℎ= the vertical distance from the
base of a slice to a Water Surface and 𝐻𝑢 = the Hu coefficient for the soil type (either
user-defined or Auto). For this case, it is defined as auto 𝐻𝑢 = 𝑐𝑜𝑠 2 𝛼 to be calculated
from the water surface where 𝛼 is water surface inclination as shown in Figure 3.9.

Since the analysis is carried out in undrained conditions excess pore water pressure is
expected apart from the initial pore pressure calculated previously. For this excess pore
pressure calculation due to a sudden increase in pore pressure within a soil due to rapidly
applied loading conditions (undrained loading) a "B-bar" method is used from Skempton
1954. It says that the change in pore pressure is assumed to be directly proportional to
the change in vertical stress.

MSc Thesis Page 56

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

∇𝑢 = 𝐵̅ ∇𝜎𝑣 (3.3)

Where, ∇𝑢 is change in excess pore pressure, 𝐵̅ is the overall pressure coefficient

(Skempton), ∇𝜎𝑣 is the change in vertical stress that may be due to the weight of added
layers of material, external loads, seismic loads, or a combination of these factors. Hence,
the final pore pressure which is used in the stability computation is equal to the Initial
Pore Pressure + Excess Pore Pressure.

Figure 3.9 Automatic Calculation of Hu coefficient (Rocscience web help, 2021)

In a conventional or traditional slope stability analysis, it is assumed that the values of

all model input parameters are precisely known. For a given slip surface, a single value
of the safety factor is calculated. This type of analysis can be referred as a deterministic
analysis. For the probabilistic analysis, it needs to define some additional input data over
the deterministic one. Since it is difficult to be precise in knowing soil properties,
probabilistic method of analysis is better in compromising this uncertainty of soil
parameters. For that matter, parameters are defined as a random variable and by assigning
statistical distribution together with the statistical parameters of the distribution (mean,
standard deviation, minimum and maximum values). By defining a "probability density

MSc Thesis Page 57

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

function" for the random variable random input data samples are generated. Then a given
slip surface has different safety factors from that probability of failure is calculated.

The values found from soil investigation are taken as mean values and the standard
deviation is calculated from the coefficient of variation value from literature as described
elsewhere in the literature section by using the following equation. Maximum and
minimum values are determined automatically.
𝜎
COV(V) = (3.4)
𝑋̅

Where, σ is standard deviation, 𝑋̅ is mean and V is coefficient of variation. a summary

is listed in Table3.3.

Table 3.4 Defined COV from different kinds of literature

S.NO. Material COV (%) Source
Cohesion 25 Cherubini 1997
1 Unit weight 6 Wang and his colleagues 2010
Friction angle 8 Harr, Kalhawy, and Duncan

Finally, a factor of safety is calculated as follows:

𝜏𝑓𝑓
𝐹𝑆 = (3.5)
𝜏

where 𝜏𝑓𝑓 is the maximum shear stress that the soil can sustain at the value of normal
stress of 𝜎𝑛 , 𝜏 is the actual shear stress applied to the soil.

In addition to factor of safety and probability of failure, the reliability index is the other
measure of slope instability in probabilistic analysis which is calculated by using the
Equation (3.6) for normal and lognormal distribution (3.7).

𝜇−1
𝛽= (3.6)
𝜎
Where, 𝛽 is normal reliability index, 𝜇 is mean factor of safety and 𝜎 is the standard
deviation of a factor of safety.

MSc Thesis Page 58

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

𝜇
ln ⌊ ⌋
√1 + 𝑉2 (3.7)
𝛽𝑙𝑛 =
√ln( 1 + 𝑉 2 )

Where, 𝛽𝑙𝑛 is lognormal reliability index, 𝜇 is the mean factor of safety and V is the
coefficient of variation.

3.13.2 RS2 V11.0

The same soil profile is modeled by using FEM-based RS2 software. By using this
software two types of analyses are carried out. The first one is the FEM method of slope
stability analysis and the second one is probabilistic slope stability analysis. All input
parameters are the same as those parameters defined in Slide however, additional
parameters are required for the analysis carried on in later software. These are elastic
modulus, poison’s ratio, and porosity.

For the finite element analysis, the following necessary steps are followed:

✓ Delineation and discretization of the domain

✓ Writing element stiffness matrix
✓ Assembling the element stiffness matrix to get global stiffness matrix
✓ Applying appropriate boundary condition
✓ Solving the equations
✓ Post processing

The delineated and discretized model is shown in Figure 3.10:

MSc Thesis Page 59

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Figure 3.10 Discretized and meshed model by RS2 Asko

Figure 3. 11 Discretized and meshed model by RS2 Kolfe 1

Quadratic triangular element with three nodes at corner and three at the middle 6 nodes
with two degrees of freedom for each node is defined. The quality of the mesh is checked
all good. For the whole model, the number of elements becomes 4929 with 19730 degrees
of freedom.

MSc Thesis Page 60

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Figure 3.12 Quadratic Triangular element

The analysis type is assumed to be a plain strain in which failures are of infinite length
in the out-of-plane direction, and therefore the strain in the out-of-plane direction is zero.

In RS2 the following are computed in a plain strain case.

✓ the major and minor in-plane principal stresses (𝜎1 and 𝜎3 )

✓ the out-of-plane principal stress (𝜎𝑍 )
✓ in-plane displacements and strains

where, 𝜎1 and 𝜎3 are computed by using the following equations.

𝜎𝑧 + 𝜎𝑥 𝜎𝑧 − 𝜎𝑥 2
𝜎1 = + √[ ] + 𝜏𝑧𝑥 2 (3.8)
2 2

𝜎𝑧 + 𝜎𝑥 𝜎𝑧 − 𝜎𝑥 2
𝜎3 = − √[ ] + 𝜏𝑧𝑥 2 (3.9)
2 2

The other important step is the material definition section. The elastic material properties
of Young’s Modulus and Poisson’s Ratio are assumed to be isotropic implies there is no
variation of these properties in every direction. When we come to the material strength
parameter, we need to consider the failure (strength) criteria and the material type (elastic
or plastic). There are different types of strength criteria in RS2. One of these is the

MSc Thesis Page 61

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Elastic/Plastic mode of failure includes Mohr-Coulomb, Hoek-Brown, Drucker-Prager,

Generalized Hoek-Brown, and Discrete function strength criteria. For this research case,
Mohr-Coulomb strength type is used for soil.

The material type by itself is defined as plastic with equal residual and peak strength
parameters which means the material is elastic-plastic. In the elastoplastic constitutive
model yield surface and potential functions are used for the derivation of constitutive
equations of the materials. In general, in the formation of elastoplasticity yield criterion
function, flow rule, Hooke’s law, Additivity postulate, and hardening rule are involved
as stated below respectively.

𝐹𝑚 (𝜎𝑖𝑗 , ҡ)= 0 (3.10)

Where 𝜎𝑖𝑗 is stress tensor, ҡ material state parameter, and m tell a number of yield surface.

𝑝 𝜕𝑄𝑚
𝑄𝑚 (𝜎𝑖𝑗 ) = 𝐶𝑜𝑛𝑠𝑡𝑎𝑛𝑡, 𝜀̇𝑖𝑗 = ∑ 𝜆̇𝑚 (3.11)
𝜕𝜎𝑖𝑗
𝑚

𝒑
it is assumed that the plastic strain increments (𝜺̇ 𝒊𝒋 ) are coaxial with the gradient of plastic

potential function 𝑄.

𝜎̇𝑖𝑗 = 𝐷𝑒 𝑖𝑗𝑘𝑙 𝜀̇ 𝑒 𝑘𝑙 (3.12)

𝜀̇𝑘𝑙 = 𝜀̇𝑒 𝑘𝑙 + 𝜀̇𝑝 𝑘𝑙 (3.13)

𝑘̇ = 𝑓(𝜀 𝑝 𝑘𝑙 ) (3.14)

Where 𝜀 𝑝 𝑘𝑙 plastic strain increment, 𝜀 𝑒 𝑘𝑙 is elastic strain increment and 𝜆̇𝑚 plastic
multiplier and equation 3.11 is called flow rule.

Hooke’s law (3.12) relates the increment of stress to the increment of elastic strains using
the elastic constitutive equations. The additivity postulate (3.13) simply sums up the
increment of elastic and plastic strain to form the increment of strain. The hardening
(3.14) controls the behavior of material after initial yielding. The expansion/shrinkage of
the yield loci is controlled by hardening/softening rules. One of the elastoplastic model
Mohr-Coulomb failure criteria in our case in terms of principal stress is given by:

MSc Thesis Page 62

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

1 1
𝐹𝑠 = (𝜎1 − 𝜎3 ) + (𝜎1 + 𝜎3 ) sin 𝜑 − 𝑐 cos 𝜑 = 0 (3.15)
2 2

The analysis is carried out by the principle of shear strength reduction factor as depicted
by the equation (3.16) as reduction of strength until a failure happens.

𝜏 𝑐 − 𝜎𝑛 tan 𝜑
= (3.16)
𝑆𝑅𝐹 𝑆𝑅𝐹

After the application of SRF, c and 𝜑 becomes:

𝑐 tan 𝜑
𝑐𝑆𝑅𝐹 = 𝑎𝑛𝑑 𝜑𝑆𝑅𝐹 = atan ( ) (3.17)
𝑆𝑅𝐹 𝑆𝑅𝐹

Finally, for the probabilistic analysis, the materials are defined as a random variable with
appropriate statistical distribution function by using the Latin hypercube method, the
same procedure followed as discussed for Slide software.

MSc Thesis Page 63

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

CHAPTER 4 RESULTS AND DISCUSSIONS

This chapter presents the major findings and discussion on results in a brief way based
on the analysis type carried on. As mentioned in the methodology section two major
types of analyses were used and these are the Limit Equilibrium Method (LEM) and
Finite Element Method (FEM) of analysis. However, the focus is on probabilistic
analysis by using the Rocscience software package for both limit equilibrium and finite
element method of analysis.

4.1 Limit Equilibrium Analysis

By using Slide, factor of safety is computed by both deterministic and probabilistic

analysis. In LEM, it is possible to get a better result depending on how we search for the
minimum slip surface. The researcher has tried all the different ways of searching
mechanisms and the best one is selected (i.e., Auto refine search). With this search
method, a probabilistic based LEM factor of safety is computed by using global
minimum analysis and overall slope analysis, and nearly the same results were found,
the results are summarized in Table 4.1 for global minimum analysis.

The global minimum analysis result is shown in Figure 4.1. Since it is a probabilistic
analysis, the expected output values are a deterministic factor of safety, mean factor of
safety, probability of failure, and reliability index for the normal and lognormal
distribution. Pictures showing the minimum slip surfaces are listed in Appendix D.

MSc Thesis Page 64

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

FS (deterministic) =1.203
FS (mean) = 1.205
PF = 0.6%
RI (normal) = 2.354
RI (lognormal) = 2.548

Figure 4.1 Global minimum analysis showing global minimum slip surface for Bishop,
Addisu gebeya 2

FS (deterministic) = 1.203
FS overall(mean) = 1.198
PF = 1.3%
RI (normal) = 2.198
RI (lognormal) = 2.368

Figure 4.2 Global minimum analysis showing all slip surfaces for Bishop, Addisu gebeya 2

Table 4.1 Factor of safeties, global minimum analysis type

Deterministic
Method of Analysis Probabilistic Analysis
Analysis
Kolfe 1
Reliability index Probability
FS FS of failure
normal lognormal (%)
Bishop simplified 1.194 1.197 1.580 1.679 5.400
GLE/Morgenstern-
1.192 1.195 1.850 1.678 5.300
Price
Janbu Simplified 1.082 1.085 0.726 0.704 22.90
Spencer 1.195 1.196 1.588 1.688 5.100

MSc Thesis Page 65

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Kolfe 2
Bishop simplified 1.355 1.359 3.148 3.619 0.200
GLE/Morgenstern-
1.354 1.357 3.134 3.601 0.200
Price
Janbu Simplified 1.212 1.216 2.054 2.222 2.400
Spencer 1.354 1.357 3.131 3.598 0.200
Kolfe 3
Bishop simplified 2.872 2.883 2.544 4.065 0.300
GLE/Morgenstern-
2.880 2.895 2.552 4.085 0.300
Price
Janbu Simplified 2.785 2.796 2.501 3.942 0.300
Spencer 2.882 2.889 2.548 4.072 0.300
Addisu gebeya 1
Bishop simplified 1.341 1.347 2.064 2.333 2.300
GLE/Morgenstern-
1.339 1.344 2.050 2.315 2.400
Price
Janbu Simplified 1.210 1.216 1.381 1.462 9.300
Spencer 1.339 1.345 2.049 2.314 2.400
Addisu gebeya 2
Bishop simplified 1.203 1.205 2.354 2.548 0.600
GLE/Morgenstern-
1.196 1.201 2.322 2.509 0.600
Price
Janbu Simplified 1.050 1.052 0.706 0.690 22.400
Spencer 1.195 1.201 2.338 2.526 0.600
Arada
Bishop simplified 2.265 2.272 3.164 4.587 0.000
GLE/Morgenstern-
2.264 2.271 3.166 4.588 0.000
Price
Janbu Simplified 2.164 2.169 2.946 4.176 0.000
Spencer 2.265 2.271 3.169 4.593 0.000
Asko
Bishop simplified 1.411 1.414 2.766 3.229 0.300
GLE/Morgenstern-
1.409 1.413 2.771 3.234 0.300
Price
Janbu Simplified 1.301 1.304 2.207 2.466 1.200
Spencer 1.411 1.414 2.777 3.242 0.200

Table 4.1 shows that the factor of safety of deterministic and probabilistic analysis are
nearly equal. This is something expected to happen if the analysis is correct. However,
after the concept of probabilistic analysis, a factor of safety is not the only measuring
criteria for failure but also other additional measuring ways called reliability index and
probability of failure values to be used.

MSc Thesis Page 66

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

The value of the reliability index is of two types (i.e., normal and lognormal) but the best
fit for this slope data is normal as indicated in appendix A for all analysis methods. For
a safe design of slope, the reliability index value should be at least 3 or above.
Nevertheless, that is attained for some slopes in this case (reliability index value, Table
4.1).

Finally, the probability of failure is the other measure of slope stability. From the last
column of Table 4.1, in all cases the failure probability is more than nil in percent except
for the case of Arada. If we take the deterministic factor of safety value of Bishop’s
Simplified method of analysis of Arada and Janbu simplified method of Addisu gebeya
2, it is 1.194 and 1.05 respectively. some one can say that this slope is safe since these
values are greater than 1 if 1 is taken as a reference for the general case. However, even
if FS>1 the probability of failure is 5.4% and 22.4% respectively. This implies that there
may a probability to fail for a given slope even if FS is greater than one. One thing to
bear in mind is that there is no value of failure which is termed as less (small) or that’s
okay even if its value is unity or less.

The Janbu simplified method gives a lesser value of factor of safety relative to the others.
However, the probability value is greater than the rest methods. This assures the inverse
proportionality between the probability of failure and the safety factor. Output results of
GM slope analysis are presented in Appendix A.

4.2 Finite Element Analysis

The second major type of analysis used for this research is the finite element method of
analysis. It is a more rigorous and advanced type than that of LEM. By using the RS2
program both the pure finite element method of analysis and probabilistic finite element
analysis are computed. The results are presented separately in the next section.

4.2.1 FEM by using RS2

One of the limitations of the limit equilibrium method is assuming the shape of the slip
surface prior to failure. Due to the SSR technique in the finite element method, no-slip
surface is assumed before failure occurs. As depicted in Figure 4.3, the critical strength
reduction factor which is the same as a factor of safety of the slope becomes 1.17 for

MSc Thesis Page 67

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Addisu gebeya 2 and decreased for others also. this indicates that more critical and weak
surfaces are possible to find in FEM than LEM. Refer appendix B for output results of
FEM analysis.

Critical SRF: 1.17

Figure 4.3 Critical SRF showing contour of the maximum shear strain, Kolfe 1

Table 4.2 Summary of factor of safety result from RS2

Deterministic Analysis
Site
FS

Kolfe 1 1.07
Kolfe 2 1.18
Kolfe 3 2.60
Addisu gebeya 1 1.34
Addisu gebeya 2 1.17
Arada 2.14
Asko 1.00

With a rigorous method, it is possible to simulate the deformed shape of elements and
the way how material moves during failure. Knowing the value of displacement value is
advantageous for minimization of risks due to the failing slope. The critical displacement
value (for SRF value of 1.17, the displacement is 0.023 m from its former position) is
shown in Figure 4.5. It illustrates the deformed shape, displacement value, and
boundaries of the model.

MSc Thesis Page 68

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Critical SRF: 1.17

Figure 4.4 Deformed shape and Boundaries, Addisu gebeya 2

Figure 4.5 Maximum total displacements at critical SRF, Addisu gebeya 2

4.2.2 Probabilistic FEM Analysis by Using RS2

Even though FEM is better in showing the most critical slip surface, the parameters used
are defined as definitive or fixed variable. Nevertheless, the uncertainty of values can be
considered by defining parameters as a random variable. By doing so, a lesser factor of
safety is found i.e., from 1.17 to 0.97.

MSc Thesis Page 69

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Mean crit. SRF: 0.97

Std. Dev. Crit. SRF: 0.0024
PF: 100.00%

Figure 4.6 Mean critical SRF, showing the contour of Maximum shear plastic strain

As shown in Figure 4.6, the probability of failure is higher than LEM. The mean critical
SRF is the average critical strength reduction factor from the entire constituent files of
the probabilistic analysis and the standard deviation of the critical strength reduction
factors from the entire constituent files of the probabilistic analysis. The safety factor
from probabilistic and FEM are much close, this is due to fewer number of parameters
defined as random variables in which the shear strength is dependent for the reason that
lack of high-speed computer for computation. A better result is expected by defining
more parameters as a random variable. Refer appendix C for output results of PFEM
analysis.

As depicted in Figure 4.7, the deformed shape and boundary deformation sever at the
toe.

Mean crit. SRF: 0.97

Std. Dev. Crit. SRF: 0.0024
PF: 100.00%

Figure 4.7 Deformed shape and boundaries, Addisu gebeya 2

MSc Thesis Page 70

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

For comparison purposes, results from all analysis methods discussed in the previous
sections are compiled in Table 4.3.

Table 4.3 Summary of findings

Deterministic
Method (Average) Probabilistic (Average)
of RI RI
Pf
Analysis Site FS FS (Average (Average
(Average)
normal) lognormal)
Kolfe 1 1.166 1.168 1.436 1.437 9.68
Kolfe 2 1.318 1.322 2.866 3.260 0.75
Kolfe 3 2.854 2.866 2.536 4.041 0.30
LEM
Addisu gebeya 1 1.341 1.342 2.063 2.333 2.3
(GM)
Addisu gebeya 2 1.161 1.165 1.930 2.060 6.05
Arada 2.400 2.245 3.111 4.486 0.00
Asko 1.383 1.386 2.615 3.040 0.50
Kolfe 1 1.070 - - - -
Kolfe 2 1.180 - - - -
Kolfe 3 2.600 - - - -
FEM Addisu gebeya 1 1.340 - - - -
Addisu gebeya 2 1.170 - - - -
Arada 2.140 - - - -
Asko 1.000 - - - -
PFEM Addisu gebeya 2 1.17 0.97 - - 100.00

MSc Thesis Page 71

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

CHAPTER 5 CONCLUSIONS AND RECOMMENDATIONS

As it is stated in the first chapter, main objective of this study was showing the reliability
of probabilistic slope stability analysis over deterministic slope stability analysis. An
effort is made to meet the stated objective and the main objective is achieved by
acknowledging all the limitations of the study. The main emphasis is given for
probabilistic analysis and good results are found in both LEM and FEM-based analysis.
Even though FEM is well-suited method of analysis that avoids many assumptions and
works by the principle of seeking out the weakest surface, probabilistic analysis is more
appreciated for its account for soil uncertainties which is the limitation of FEM.

5.1 Conclusions

The following major conclusions are drawn from this research:

✓ a factor of safety greater than unity cannot guarantee the analysis to be termed as
safe because a probability of failure greater than zero is shown for FS greater than
1.
✓ FEM lies on safe side than that of both deterministic and probabilistic LEM of
analysis and finally,
✓ A finite element-based probabilistic analysis is more reliable than that of
deterministic FEM and the traditional (conventional) probabilistic method of slope
stability analysis.

5.2 Recommendation

Despite the promising results found, there are few points to be taken as a recommendation
for further investigation on the same or related topic. The number of samples used for
finite element probabilistic analysis is 70 and all parameters are not defined as random
variables due to a lack of high-performance computer if more random variables are
defined.

MSc Thesis Page 72

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

References

AASHTO. (2012). AASHTO LRFD BRIDGE DESIGN SPECIFICATIONS. Washington,

DC, US.

Abramson, L. W., Thomas S. Lee, S. L., Sharma, S., & Glenn, M. B. (2001). Slope
Stability and Stabilization Methods (2nd ed.). John Wiley and Sons, Inc.

ACI Commitee 318. (2014). Building Code Requirements for Structural Concrete (ACI
318-14) AND Commentary on Building Code Requirements for Structural
Concrete (ACI 318R-14). American Concrete Institure.

Akalu, M. (2017). Engineering Characteristics of Red Soil: A Case from Western Addis
Ababa Ethiopia. Addis Ababa.

AKBAŞ, B. (2015). Probabilistic analysis using limit equilibrium, Finite element and
random finite element. Middle East Technical University.

Alamanis, N. (2017). Uncertainties and optimization in geotechnical engineering.

American Scientific Research Journal for Engineering, Technology, and Sciences
(ASRJETS) Volume 38, 92-111.

Allahverdizadeh, p., ASCE, S. M., Griffiths, D. V., Fenton, G. A., & ASCE, M. (2015).
The random finite element method (RFEM) in probabilistic slope stability
analysis with consideration of spatial variability of soil properties. (pp. 1946-
1955). San Antonio, Texas: ASCE.
doi:https://doi.org/10.1061/9780784479087.178

Aruna, M. (2008). Shear Behavior of Concrete Beams Reinforced with High

Performance Steel Shear Reinforcement.

Aryal, K. P. (2006). Slope stability evaluations by limit equilibrium and finite element
methods. Trondheim, Norway: Norwegian University of Science and
Technology.

ASCE-ACI Committee 445. (1998, December). Recent Approaches to Shear Design of

Structural Concrete. American Society of Civil Engineers.

MSc Thesis Page 73

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Baecher , G. B., & Christian , J. T. (2005). Reliability and statistics in geotechnical

engineering. Chichester, West Sussex, England: ” John Wiley & Sons Inc.

Bentz, C. E., Vecchio, F. J., & Collins, M. P. (2006). Simplified Modified Compression
Field Theory for Calculating Shear Strength of Reinforced Concrete Elements.
ACI Structural Journal(July-August).

Bishop A.W. (1955). The Uses of the Slip Circle in the Stability Analysis of Slopes.
Geotechnique, 7-17.

Bowles, J. E. (1997). Fundation analysis and design. The McGraw-Hill Companies,Inc.

Chok, Y. H., Jaksa, M. B., Griffiths, D. V., Fenton, G. A., & Kaggwa, W. S. (2015).
probablistic analiysis of spatial variable c-phii slope. Australian Geomechanics
journal, 17-27.

Civil Seek. (2018, July 18). Slope Failure; its Types, Causes, Technical Terms. Retrieved
February 11, 2021, from civil seek: https://civilseek.com/slope-failure/

Climate- Ethiopia. (n.d.). Retrieved from Climate to Travel .

Collins, M. P., Bentz, E. C., Sherwood, E. G., & Xie, L. (2007). AN ADEQUATE
THEORY FOR THE SHEAR STRENGTH OF REINFORCED CONCRETE
STRUCTURES.

Davidson, A. (1983). Reconnaissance Geology and Geochemistry of Parts of Ilubabor

Kefa,Bull. No2. Ottawa. Ethiopia.

Derbie, M. (2019). A comparative study of deterministic and probabilistic approaches of

slope stability analysis,A Thesis in Geotechnical Engineering. Addis Ababa:
Addis Ababa University.

Duncan, J. M. (1996). State of the art: Limit equilibrium and finite-element analysis of
slopes. Journal of Geotechnical Engineering, 122(7), 577-596.
doi:https://doi.org/10.1061/(ASCE)0733-9410(1996)122:7(577)

Duncan, J. M., Member, H., & ASCE. (2000). Factors of safety and reliability in
geotechnical engineering. Journal of geotechnical and geoenvironmental
engineering, 126(4), 307-316.

MSc Thesis Page 74

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Eckhardt, R. (1987). Stan Ulam, John von Neumann, and the Monte Carlo method. Los
Alamos Science,Special Issue, 131-137.

El-Ramly, H., Morgenstern, N. R., & Cruden, D. M. (2002). Probabilistic slope stability
analysis for practice. Canadian Geotechnical Journal, 39(3), 665-683. doi:DOI:
10.1139/t02-034

El-Ramly, H., Morgenstern, N. R., & Cruden, D. M. (2002). Probabilistic slope stability
analysis for practice. Canadian Geotechnical Journal, 39(3), 665-683. doi:DOI:
10.1139/t02-034

El-Ramly, H., Morgenstern, N. R., & Cruden, D. M. (2005). Probabilistic slope stability
analysis for practice. Geotechnique, 55(1), 77-84.
doi:https://doi.org/10.1139/t02-034

(2015). ES EN 1998-1:2015 Design of Structures for earthquake resisistance .

ETİZ, M. C. (2019). Comparison of Limit Equilibrium Method and 2-D, 3-D Finite
Element Slope Stability Models: A case study on the slope in Akapinar district, in
Ankara. Ankara: middle east technical university.

European Standard. (1992). Eurocode 2: Design of Concrete Structures. European

Commitee for Standardization.

Fellenius, W. (1936). Calculation of stability of earth dams. Transactions 2nd

international congress on large dams. International Commission on Large Dams.
Washington D.C.

Fenton, G. A. (1990). Simulation of random fields via local average Subdivision. J

Geotech Geoenviron, 1733-1749.

Fenton, G. A., & Vanmarcke, E. H. (1990). Simulation of random fields via local average
subdivision. journal of engineering mechanics, 116(8), 1733-1749.
doi:https://doi.org/10.1061/(ASCE)0733-9399(1990)116:8(1733)

Fishman, G. (1995). Monte Carlo: Concepts, algorithms, and applications.

Frank, J. V. (2000, January - February). Analysis of shear-critical reinforced concrete

beams. ACI Structural Journal, 97.

MSc Thesis Page 75

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Geology of Ethiopia. (1998). Retrieved from

https://sites.google.com/site/linkstogeologyofethiopia/map/geological-map-of-
addis-ababa

Ghoneim, M., & EL-Mihilmy, M. (2008). Design of Reinforced Concrete Structures (2nd
edition ed.). Egypt: Cairo University.

Griffiths , D. V., & Fenton, G. A. (2004). Probabilistic slope stability analysis by finite
elements. Journal of Geotechnical and Geoenvironmental Engineering, 130(5),
507–518. doi:https://doi.org/10.1061/(ASCE)1090-0241(2004)130:5(507)

Griffiths , D. V., & Fenton, G. A. (2004). Probabilistic slope stability analysis by finite
elements. Journal of Geotechnical and Geoenvironmental Engineering, 130(5),
507–518. doi:https://doi.org/10.1061/(ASCE)1090-0241(2004)130:5(507)

Griffiths, D. V., Fenton, G. A., & Denavit, M. D. (2007). Traditional and advanced
probabilistic slope stability analysis. doi:https://doi.org/10.1061/40914(233)19

Griffiths, D. V., & Fenton, G. A. (1993). Seepage beneath water retaining structures
founded on spatially random soil. Geotechnique, 43(4), 577–587.
doi:https://doi.org/10.1680/geot.1993.43.4.577

Griffiths, D. V., & Fenton, G. A. (2000). Influence of soil strength spatial variability on
the stability of an undrained clay slope by finite elements. Proceeding of
GeoDenver 2000 (pp. 184–193). Denver, Colorado, United States: ASCE.
doi:https://doi.org/10.1061/40512(289)14

Griffiths, D., Huang, J., & Fenton, G. (2015). Probabilistic slope stability analysis using
RFEM with non-stationary random fields., (pp. 704-709). doi:doi:10.3233/978-
1-61499-580-7-704

Haile Sellasie, g. a. (1989). The Addis Ababa-Nazareth Volcanics A Miocene-

Pleistocene Volcanic Succession in Ethiopian Rift. Ethiopian Journal of Science.

Janbu, N. (1954a). Stability analysis of slopes with dimensionless parameters, Thesis for
the doctor of science in the field of civil engineering. Harvard University Soil
Mechanics Series.

MSc Thesis Page 76

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Le, T. M., Sanchez, M., Gallipoli, D., & Wheeler, S. (2014). Probabilistic modelling of
auto-correlation characteristics of heterogeneous slopes. Geomechanics and
Geoengineering, 10(2), 95-108.

Lukas, M. (2010). A STUDY ON THE EFFECT OF REMOLDING ON THE

MECHANICAL BEHAVIOR OF ADDIS ABABA RED CLAY SOIL. Addis Ababa.

Morgenstern, N. R., & Price, V. E. (1965). The analysis of the stability of general slip
surface. Geotechnique, 15(1), 79-93.
doi:https://doi.org/10.1680/geot.1965.15.1.79

Murthy, V. s. (2002). Principles and practices soil mechanics and foundation

engineering. New York: Taylor & Francis Inc.

Nakamura, A. (2008). Embankment basal stability analysis using shear strength

reduction finite element method. 10th International, Symposium, On landslides
and Engineered slopes (pp. 851-856). Boca Raton: CRC Press,.

Nilson, A. H., Darwin, D., & Dolan, W. C. (2010). Design of Concrete Structures (14th
edition ed.). New York, US: Mc Graw Hill Companies.

Oguz, E. A., Yalcin, Y., & Huvaj, a. N. (2017). Probabilistic slope stability analyses:
Effects of the coefficient of variation and the cross-correlation of shear strength
parameters. Geotechnical Frontiers 2017 GSP 278 (pp. 363-371). ASCE.

Oğuz, N. H. (2018). Probabilistic Slope Stability Analysis. Sakarya University Journal

of Science, 1458-1465.

Peter, M. (1985). Truss Models in Detailing.

Phoon, k. k., & Kulhawy, F. H. (1999). Evaluation of geotechnical property variability.

Canadian Geotechnical Journal, 36(4), 625-639. doi:https://doi.org/10.1139/t99-
039

Phoon, K. K., & Kulhway, F. H. (1999). Characterization of geotechnical variability.

Canadian Geotechnical Journal, 36(4), 612–624.
doi:https://doi.org/10.1139/t99-038

Rocscience. (2021). products. Retrieved from Rocscience:

https://www.rocscience.com/software/rs2

MSc Thesis Page 77

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Rocscience web help. (2021). Slide program overview. Retrieved from C:\Program Files

(x86)\Rocscience\Slide 6.0\WebHelp\Slide.htm

Salunkhe, D. P., Bartakke, R. N., Chvan, G., & Kothavale, P. R. (2017). An Overview

on Methods for Slope Stability Analysis. International Journal of Engineering

Research & Technology (IJERT), 6(3), 528-535.

SG Consulting Engineers PLC. (2016). Detailed Geotechnical Investigation Report on

the Landslide Problem along Jimma-Mizan Road.

Sharma, R. K. (2016). Reliability analysis of slope stability using monte carlo simulation

and comparison with deterministic analysis. International Conference on

Computational Methods(ICCM).

Spencer, E. (1967). A method of analysis of the stability of embankments assuming

parallel inter-slice forces. Geotechnique, 17(1), 11-26.

doi:https://doi.org/10.1680/geot.1967.17.1.11

Stanford University. (2014, April). Some Useful Numbers.

Teferra A. and Leikun M. (1999). Soil Mechanics. Addis Ababa: Adiss Ababa

University.

Terzaghi, k. (1950). Mechanism of landslides. Application of geology to engineering

practice. New York: Geological Society of America.

Transportation Reasearch Board Executive Commitee . (2005). Simplified Shear Design

of Structural Concrete Members. NCHRP, NCHRP REPORT 549.

Weight, K. J., & Macgregor, J. G. (2012). Reinforced Concrete Mechanics & Design (6th

edition ed.). New Jersey, US: Pearson Education.

MSc Thesis Page 78

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Woldearegay, K. (2013). Review of the occurrences and influencing factors of landslides

in the highlands of Ethiopia: With implications for infrastructural development.

Momona Ethiopian Journal of Science (MEJS) v5, 3-31.

Zakaria, M., Ueda, T., Wu, Z., & Meng, L. (2009). Experimental Investigation on Shear

Cracking Behavior in Reinforced Concrete Beams with Shear Reinforcement.

Zhang, Y., Ji, E., & Xu, W. (2020). In Probabilistic slope stability analysis for

embankment dams (pp. 1-19). Nanjing, China.

MSc Thesis Page 79

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

APPENDIX A

Slide Analysis Information

SLIDE - An Interactive Slope Stability Program

Project Summary
• File Name: Addisu gebeya 2.slim
• Slide Modeler Version: 6.02
• Project Title: SLIDE - An Interactive Slope Stability Program
• Date Created: 3/24/2022, 11:40:57 AM

General Settings
• Units of Measurement: Metric Units
• Time Units: days
• Permeability Units: meters/second
• Failure Direction: Right to Left
• Data Output: Standard
• Maximum Material Properties: 20
• Maximum Support Properties: 20

Analysis Options
Analysis Methods Used

• Bishop simplified
• GLE/Morgenstern-Price with interslice force function: Half Sine
• Janbu simplified
• Janbu corrected
• Spencer

• Number of slices: 25
• Tolerance: 0.005
• Maximum number of iterations: 50
• Check malpha < 0.2: Yes
• Initial trial value of FS: 1
• Steffensen Iteration: Yes

Groundwater Analysis
• Groundwater Method: Water Surfaces
• Pore Fluid Unit Weight: 9.81 kN/m3

MSc Thesis Page 80

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

• Advanced Groundwater Method: Excess Pore Pressure

Random Numbers
• Pseudo-random Seed: 10116
• Random Number Generation Method: Park and Miller v.3

Surface Options
• Surface Type: Circular
• Search Method: Auto Refine Search
• Divisions along slope: 10
• Circles per division: 10
• Number of iterations: 10
• Divisions to use in next iteration: 50%
• Composite Surfaces: Disabled
• Minimum Elevation: Not Defined
• Minimum Depth: Not Defined

Loading
• 4 Distributed Loads present

Distributed Load 1

• Distribution: Constant
• Magnitude [kPa]: 96.66
• Orientation: Normal to boundary
• Creates Excess Pore Pressure: No

Distributed Load 2

• Distribution: Constant
• Magnitude [kPa]: 9.3
• Orientation: Normal to boundary
• Creates Excess Pore Pressure: No

Distributed Load 3

• Distribution: Constant
• Magnitude [kPa]: 9.3
• Orientation: Normal to boundary
• Creates Excess Pore Pressure: No

MSc Thesis Page 81

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Distributed Load 4

• Distribution: Constant
• Magnitude [kPa]: 96.66
• Orientation: Normal to boundary
• Creates Excess Pore Pressure: No

Material Properties

Property Red Clay-Addisu Gebeya 2

Color ___
Strength Type Mohr-Coulomb
Unit Weight [kN/m3] 19.5
Cohesion [kPa] 15.3
Friction Angle [deg] 21.3
Water Surface Water Table
Hu Value Automatically Calculated
Material Weight Causes Excess Pore Pressure
B_bar value 0.18

Probabilistic Analysis Input

General Settings

• Sensitivity Analysis: On
• Probabilistic Analysis: On
• Sampling Method: Latin-Hypercube
• Number of Samples: 1000
• Analysis Type: Global Minimum

Variables

Standard
Material Property Distribution Mean Min Max
Deviation
Red Clay-Addisu
Cohesion Normal 15.3 3.825 26.775 3.825
Gebeya 2
Red Clay-Addisu
Phi Normal 21.3 16.188 26.412 1.704
Gebeya 2
Red Clay-Addisu Unit
Normal 19.5 15.99 23.01 1.17
Gebeya 2 Weight

MSc Thesis Page 82

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Correlation Coefficients

Material Correlation
Red Clay-Addisu Gebeya 2 -0.5

Global Minimums

Method: bishop simplified

• FS: 1.202620
• Center: 4.727, 26.038
• Radius: 26.483
• Left Slip Surface Endpoint: -0.108, 0.000
• Right Slip Surface Endpoint: 26.809, 11.418
• Resisting Moment=43736.8 kN-m
• Driving Moment=36368 kN-m
• Total Slice Area=132.185 m2

Method: janbu simplified

• FS: 1.050110
• Center: 7.258, 18.655
• Radius: 20.723
• Left Slip Surface Endpoint: -1.766, 0.000
• Right Slip Surface Endpoint: 26.677, 11.418
• Resisting Horizontal Force=1618.01 kN
• Driving Horizontal Force=1540.8 kN
• Total Slice Area=178.123 m2

Method: janbu corrected

• FS: 1.130140
• Center: 7.349, 18.603
• Radius: 20.620
• Left Slip Surface Endpoint: -1.545, 0.000
• Right Slip Surface Endpoint: 26.677, 11.418
• Resisting Horizontal Force=1732.89 kN
• Driving Horizontal Force=1533.34 kN
• Total Slice Area=177.086 m2

Method: spencer

• FS: 1.195030
• Center: 4.891, 26.323
• Radius: 26.836
• Left Slip Surface Endpoint: -0.329, 0.000
• Right Slip Surface Endpoint: 27.207, 11.418

MSc Thesis Page 83

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

• Resisting Moment=45602.8 kN-m

• Driving Moment=38160.2 kN-m
• Resisting Horizontal Force=1486.66 kN
• Driving Horizontal Force=1244.03 kN
• Total Slice Area=137.318 m2

Method: gle/morgenstern-price

• FS: 1.196050
• Center: 4.727, 26.038
• Radius: 26.483
• Left Slip Surface Endpoint: -0.108, 0.000
• Right Slip Surface Endpoint: 26.809, 11.418
• Resisting Moment=43497.9 kN-m
• Driving Moment=36368 kN-m
• Resisting Horizontal Force=1431.67 kN
• Driving Horizontal Force=1197 kN
• Total Slice Area=132.185 m2

Valid / Invalid Surfaces

Method: bishop simplified

• Number of Valid Surfaces: 2393

• Number of Invalid Surfaces: 2

Error Codes:

o Error Code -112 reported for 2 surfaces

Method: janbu simplified

• Number of Valid Surfaces: 2362

• Number of Invalid Surfaces: 33

Error Codes:

o Error Code -108 reported for 33 surfaces

Method: janbu corrected

• Number of Valid Surfaces: 2362

• Number of Invalid Surfaces: 33

MSc Thesis Page 84

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Error Codes:

o Error Code -108 reported for 33 surfaces

Method: spencer

• Number of Valid Surfaces: 2233

• Number of Invalid Surfaces: 162

Error Codes:

o Error Code -108 reported for 89 surfaces

o Error Code -111 reported for 70 surfaces
o Error Code -112 reported for 3 surfaces

Method: gle/morgenstern-price

• Number of Valid Surfaces: 2280

• Number of Invalid Surfaces: 115

Error Codes:

o Error Code -108 reported for 44 surfaces

o Error Code -111 reported for 69 surfaces
o Error Code -112 reported for 2 surfaces

Error Codes

The following errors were encountered during the computation:

o -108 = Total driving moment or total driving force < 0.1. This is to limit the calculation
of extremely high safety factors if the driving force is very small (0.1 is an arbitrary
number).
o -111 = safety factor equation did not converge
o -112 = The coefficient M-Alpha = cos(alpha)(1+tan(alpha)tan(phi)/F) < 0.2 for the final
iteration of the safety factor calculation. This screens out some slip surfaces which
may not be valid in the context of the analysis, in particular, deep seated slip surfaces
with many high negative base angle slices in the passive zone

MSc Thesis Page 85

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional Probabilistic Method in Slope Stability Analysis

Slice Data
• Global Minimum Query (bishop simplified) - Safety Factor: 1.20262
Slice Base Base Shear Shear Base Pore Effective
Width Weight Base
Numbe Cohesion Friction Angle Stress Strength Normal Stress Pressure Normal Stress
[m] [kN] Material
r [kPa] [degrees] [kPa] [kPa] [kPa] [kPa] [kPa]
1 1.0767 8.02413 Red Clay-Addisu Gebeya 2 15.3 21.3 15.9904 19.2303 10.0807 0 10.0807
2 1.0767 26.4375 Red Clay-Addisu Gebeya 2 15.3 21.3 21.5376 25.9015 27.1914 0 27.1914
3 1.0767 43.9872 Red Clay-Addisu Gebeya 2 15.3 21.3 26.6686 32.0722 43.0184 0 43.0184
4 1.0767 60.6078 Red Clay-Addisu Gebeya 2 15.3 21.3 31.3812 37.7397 57.5547 0 57.5547
5 1.0767 76.3061 Red Clay-Addisu Gebeya 2 15.3 21.3 35.6937 42.9259 70.8568 0 70.8568
6 1.0767 91.0843 Red Clay-Addisu Gebeya 2 15.3 21.3 39.3665 47.343 82.9796 0.793523 82.186
7 1.0767 104.94 Red Clay-Addisu Gebeya 2 15.3 21.3 42.3892 50.9781 93.9919 2.48185 91.5101
8 1.0767 117.867 Red Clay-Addisu Gebeya 2 15.3 21.3 45.1885 54.3446 103.902 3.75773 100.144
9 1.0767 127.259 Red Clay-Addisu Gebeya 2 15.3 21.3 47.0258 56.5542 110.427 4.61592 105.811
10 1.0767 124.365 Red Clay-Addisu Gebeya 2 15.3 21.3 45.4574 54.668 106.053 5.07905 100.974
11 1.0767 123.179 Red Clay-Addisu Gebeya 2 15.3 21.3 44.4871 53.5011 103.199 5.21899 97.9803
12 1.0767 131.644 Red Clay-Addisu Gebeya 2 15.3 21.3 46.3057 55.6882 108.504 4.9138 103.591
13 1.0767 139.637 Red Clay-Addisu Gebeya 2 15.3 21.3 48.0584 57.796 113.144 4.14768 108.996
14 1.0767 146.541 Red Clay-Addisu Gebeya 2 15.3 21.3 49.5801 59.626 116.591 2.90119 113.69
15 1.0767 152.302 Red Clay-Addisu Gebeya 2 15.3 21.3 50.8635 61.1695 118.8 1.15059 117.649
16 1.0767 156.854 Red Clay-Addisu Gebeya 2 15.3 21.3 51.5829 62.0346 119.868 0 119.868
17 1.0767 160.116 Red Clay-Addisu Gebeya 2 15.3 21.3 51.5802 62.0314 119.86 0 119.86
18 1.0767 161.218 Red Clay-Addisu Gebeya 2 15.3 21.3 50.9694 61.2968 117.976 0 117.976
19 1.0767 149.877 Red Clay-Addisu Gebeya 2 15.3 21.3 47.2395 56.8112 106.47 0 106.47
20 1.0767 133.33 Red Clay-Addisu Gebeya 2 15.3 21.3 51.4957 61.9298 119.599 0 119.599
21 1.0767 114.923 Red Clay-Addisu Gebeya 2 15.3 21.3 63.913 76.8631 157.901 0 157.901
22 1.0767 94.3977 Red Clay-Addisu Gebeya 2 15.3 21.3 57.6135 69.2871 138.47 0 138.47

MSc Thesis Page 86

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional Probabilistic Method in Slope Stability Analysis

23 1.0767 71.4064 Red Clay-Addisu Gebeya 2 15.3 21.3 50.8476 61.1503 117.6 0 117.6
24 1.0767 45.4613 Red Clay-Addisu Gebeya 2 15.3 21.3 43.5548 52.3799 95.1052 0 95.1052
25 1.0767 15.8392 Red Clay-Addisu Gebeya 2 15.3 21.3 35.6549 42.8793 70.7374 0 70.7374

• Global Minimum Query (janbu simplified) - Safety Factor: 1.05011

Base Base Shear Shear Base Pore Effective
Slice Width Weight Base
Cohesion Friction Angle Stress Strength Normal Stress Pressure Normal Stress
Number [m] [kN] Material
[kPa] [degrees] [kPa] [kPa] [kPa] [kPa] [kPa]
1 1.1377 5.64312 Red Clay-Addisu Gebeya 2 15.3 21.3 19.6778 20.6639 13.7576 0 13.7576
2 1.1377 17.759 Red Clay-Addisu Gebeya 2 15.3 21.3 23.6856 24.8725 24.5522 0 24.5522
3 1.1377 40.8975 Red Clay-Addisu Gebeya 2 15.3 21.3 31.5809 33.1634 45.8173 0 45.8173
4 1.1377 65.0324 Red Clay-Addisu Gebeya 2 15.3 21.3 38.4099 40.3346 66.7991 2.58869 64.2104
5 1.1377 87.6447 Red Clay-Addisu Gebeya 2 15.3 21.3 43.5368 45.7184 85.3914 7.37243 78.019
6 1.1377 108.791 Red Clay-Addisu Gebeya 2 15.3 21.3 48.1997 50.615 102.119 11.5403 90.5787
7 1.1377 128.51 Red Clay-Addisu Gebeya 2 15.3 21.3 52.4341 55.0616 117.092 15.109 101.983
8 1.1377 146.828 Red Clay-Addisu Gebeya 2 15.3 21.3 56.2657 59.0852 130.392 18.0891 112.303
9 1.1377 163.757 Red Clay-Addisu Gebeya 2 15.3 21.3 59.713 62.7052 142.074 20.4857 121.588
10 1.1377 176.456 Red Clay-Addisu Gebeya 2 15.3 21.3 61.8901 64.9914 149.75 22.2982 127.452
11 1.1377 175.086 Red Clay-Addisu Gebeya 2 15.3 21.3 59.8398 62.8384 145.371 23.4413 121.929
12 1.1377 176.485 Red Clay-Addisu Gebeya 2 15.3 21.3 58.8129 61.76 143.373 24.2093 119.164
13 1.1377 187.439 Red Clay-Addisu Gebeya 2 15.3 21.3 60.8404 63.8891 148.987 24.3627 124.624
14 1.1377 197.14 Red Clay-Addisu Gebeya 2 15.3 21.3 62.5783 65.7141 153.182 23.8768 129.306
15 1.1377 205.229 Red Clay-Addisu Gebeya 2 15.3 21.3 63.9304 67.134 155.667 22.7189 132.948
16 1.1377 211.605 Red Clay-Addisu Gebeya 2 15.3 21.3 64.8786 68.1297 156.347 20.8455 135.501
17 1.1377 216.13 Red Clay-Addisu Gebeya 2 15.3 21.3 65.3965 68.6735 155.096 18.2002 136.896
18 1.1377 218.627 Red Clay-Addisu Gebeya 2 15.3 21.3 65.4476 68.7272 151.742 14.7083 137.033
19 1.1377 213.279 Red Clay-Addisu Gebeya 2 15.3 21.3 63.5415 66.7256 142.17 10.2706 131.9
20 1.1377 194.203 Red Clay-Addisu Gebeya 2 15.3 21.3 58.2247 61.1423 122.783 5.20379 117.58

MSc Thesis Page 87

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional Probabilistic Method in Slope Stability Analysis

21 1.1377 171.769 Red Clay-Addisu Gebeya 2 15.3 21.3 79.2139 83.1833 174.112 0 174.112
22 1.1377 145.605 Red Clay-Addisu Gebeya 2 15.3 21.3 71.6677 75.259 153.787 0 153.787
23 1.1377 114.624 Red Clay-Addisu Gebeya 2 15.3 21.3 61.0387 64.0974 125.159 0 125.159
24 1.1377 76.8721 Red Clay-Addisu Gebeya 2 15.3 21.3 48.933 51.385 92.5533 0 92.5533
25 1.1377 27.9924 Red Clay-Addisu Gebeya 2 15.3 21.3 34.5762 36.3088 53.8848 0 53.8848

• Global Minimum Query (janbu corrected) - Safety Factor: 1.13014

Base Base Shear Shear Base Pore Effective
Slice Width Weight Base
Cohesion Friction Angle Stress Strength Normal Stress Pressure Normal Stress
Number [m] [kN] Material
[kPa] [degrees] [kPa] [kPa] [kPa] [kPa] [kPa]
1 1.12886 5.4895 Red Clay-Addisu Gebeya 2 15.3 21.3 18.2009 20.5696 13.5157 0 13.5157
2 1.12886 18.953 Red Clay-Addisu Gebeya 2 15.3 21.3 22.4355 25.3553 25.7904 0 25.7904
3 1.12886 43.0369 Red Clay-Addisu Gebeya 2 15.3 21.3 30.1394 34.0617 48.1213 0 48.1213
4 1.12886 66.6956 Red Clay-Addisu Gebeya 2 15.3 21.3 36.3124 41.0381 68.734 2.71922 66.0148
5 1.12886 88.8637 Red Clay-Addisu Gebeya 2 15.3 21.3 40.9993 46.335 87.0267 7.4259 79.6008
6 1.12886 109.595 Red Clay-Addisu Gebeya 2 15.3 21.3 45.2631 51.1536 103.484 11.5245 91.9596
7 1.12886 128.928 Red Clay-Addisu Gebeya 2 15.3 21.3 49.1351 55.5295 118.214 15.031 103.183
8 1.12886 146.886 Red Clay-Addisu Gebeya 2 15.3 21.3 52.6384 59.4888 131.294 17.9556 113.338
9 1.12886 163.48 Red Clay-Addisu Gebeya 2 15.3 21.3 55.7895 63.0499 142.775 20.3028 122.472
10 1.12886 174.608 Red Clay-Addisu Gebeya 2 15.3 21.3 57.386 64.8542 149.172 22.0721 127.1
11 1.12886 172.455 Red Clay-Addisu Gebeya 2 15.3 21.3 55.2631 62.455 144.143 23.1964 120.946
12 1.12886 175.353 Red Clay-Addisu Gebeya 2 15.3 21.3 54.7579 61.8841 143.414 23.9317 119.482
13 1.12886 186.306 Red Clay-Addisu Gebeya 2 15.3 21.3 56.671 64.0462 149.086 24.058 125.028
14 1.12886 195.796 Red Clay-Addisu Gebeya 2 15.3 21.3 58.2525 65.8335 153.163 23.5507 129.612
15 1.12886 203.701 Red Clay-Addisu Gebeya 2 15.3 21.3 59.4789 67.2195 155.544 22.377 133.167
16 1.12886 209.92 Red Clay-Addisu Gebeya 2 15.3 21.3 60.3329 68.1846 156.135 20.4937 135.642
17 1.12886 214.317 Red Clay-Addisu Gebeya 2 15.3 21.3 60.7901 68.7013 154.812 17.8446 136.968
18 1.12886 216.718 Red Clay-Addisu Gebeya 2 15.3 21.3 60.8166 68.7313 151.399 14.3555 137.044

MSc Thesis Page 88

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional Probabilistic Method in Slope Stability Analysis

19 1.12886 210.646 Red Clay-Addisu Gebeya 2 15.3 21.3 58.8585 66.5184 141.296 9.92754 131.369
20 1.12886 191.659 Red Clay-Addisu Gebeya 2 15.3 21.3 53.8911 60.9045 121.91 4.94057 116.969
21 1.12886 169.509 Red Clay-Addisu Gebeya 2 15.3 21.3 74.3855 84.066 176.376 0 176.376
22 1.12886 143.683 Red Clay-Addisu Gebeya 2 15.3 21.3 66.3737 75.0116 153.152 0 153.152
23 1.12886 113.109 Red Clay-Addisu Gebeya 2 15.3 21.3 56.5453 63.9041 124.663 0 124.663
24 1.12886 75.8566 Red Clay-Addisu Gebeya 2 15.3 21.3 45.35 51.2519 92.212 0 92.212
25 1.12886 27.6232 Red Clay-Addisu Gebeya 2 15.3 21.3 32.0687 36.2421 53.7135 0 53.7135

• Global Minimum Query (spencer) - Safety Factor: 1.19503

Base Base Shear Shear Base Pore Effective
Slice Width Weight Base
Cohesion Friction Angle Stress Strength Normal Stress Pressure Normal Stress
Number [m] [kN] Material
[kPa] [degrees] [kPa] [kPa] [kPa] [kPa] [kPa]
1 1.10146 6.00992 Red Clay-Addisu Gebeya 2 15.3 21.3 18.4663 22.0678 17.3584 0 17.3584
2 1.10146 24.9297 Red Clay-Addisu Gebeya 2 15.3 21.3 25.2991 30.2332 38.3016 0 38.3016
3 1.10146 43.5646 Red Clay-Addisu Gebeya 2 15.3 21.3 31.526 37.6745 57.3877 0 57.3877
4 1.10146 61.215 Red Clay-Addisu Gebeya 2 15.3 21.3 36.9612 44.1698 74.0473 0 74.0473
5 1.10146 77.8895 Red Clay-Addisu Gebeya 2 15.3 21.3 41.6855 49.8154 88.5276 0 88.5276
6 1.10146 93.5919 Red Clay-Addisu Gebeya 2 15.3 21.3 45.2721 54.1015 100.845 1.32462 99.5208
7 1.10146 108.321 Red Clay-Addisu Gebeya 2 15.3 21.3 48.1 57.481 111.339 3.15098 108.188
8 1.10146 122.07 Red Clay-Addisu Gebeya 2 15.3 21.3 50.5546 60.4143 120.264 4.55231 115.712
9 1.10146 132.221 Red Clay-Addisu Gebeya 2 15.3 21.3 51.8782 61.996 125.293 5.52367 119.769
10 1.10146 129.479 Red Clay-Addisu Gebeya 2 15.3 21.3 49.4468 59.0904 118.4 6.08332 112.317
11 1.10146 128.801 Red Clay-Addisu Gebeya 2 15.3 21.3 47.8189 57.145 113.642 6.3151 107.327
12 1.10146 137.987 Red Clay-Addisu Gebeya 2 15.3 21.3 49.1082 58.6858 117.369 6.08994 111.279
13 1.10146 146.551 Red Clay-Addisu Gebeya 2 15.3 21.3 50.2612 60.0637 120.205 5.39201 114.813
14 1.10146 153.973 Red Clay-Addisu Gebeya 2 15.3 21.3 51.1567 61.1338 121.759 4.20175 117.557
15 1.10146 160.197 Red Clay-Addisu Gebeya 2 15.3 21.3 51.8057 61.9094 122.042 2.49524 119.547
16 1.10146 165.156 Red Clay-Addisu Gebeya 2 15.3 21.3 52.2174 62.4014 121.052 0.243239 120.809

MSc Thesis Page 89

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional Probabilistic Method in Slope Stability Analysis

17 1.10146 168.766 Red Clay-Addisu Gebeya 2 15.3 21.3 51.585 61.6456 118.87 0 118.87
18 1.10146 168.808 Red Clay-Addisu Gebeya 2 15.3 21.3 50.0093 59.7626 114.04 0 114.04
19 1.10146 155.309 Red Clay-Addisu Gebeya 2 15.3 21.3 45.4907 54.3628 100.191 0 100.191
20 1.10146 138.198 Red Clay-Addisu Gebeya 2 15.3 21.3 54.3655 64.9684 127.393 0 127.393
21 1.10146 119.145 Red Clay-Addisu Gebeya 2 15.3 21.3 58.433 69.8292 139.86 0 139.86
22 1.10146 97.8848 Red Clay-Addisu Gebeya 2 15.3 21.3 52.0522 62.2039 120.303 0 120.303
23 1.10146 74.0562 Red Clay-Addisu Gebeya 2 15.3 21.3 45.4882 54.3598 100.183 0 100.183
24 1.10146 47.154 Red Clay-Addisu Gebeya 2 15.3 21.3 38.7218 46.2737 79.4435 0 79.4435
25 1.10146 16.4301 Red Clay-Addisu Gebeya 2 15.3 21.3 31.7423 37.933 58.0505 0 58.0505

• Global Minimum Query (gle/morgenstern-price) - Safety Factor: 1.19605

Base Base Shear Shear Base Pore Effective
Slice Width Weight Base
Cohesion Friction Angle Stress Strength Normal Stress Pressure Normal Stress
Number [m] [kN] Material
[kPa] [degrees] [kPa] [kPa] [kPa] [kPa] [kPa]
1 1.0767 8.02413 Red Clay-Addisu Gebeya 2 15.3 21.3 16.5197 19.7583 11.4351 0 11.4351
2 1.0767 26.4375 Red Clay-Addisu Gebeya 2 15.3 21.3 23.3523 27.9305 32.3956 0 32.3956
3 1.0767 43.9872 Red Clay-Addisu Gebeya 2 15.3 21.3 30.1516 36.0628 53.2539 0 53.2539
4 1.0767 60.6078 Red Clay-Addisu Gebeya 2 15.3 21.3 36.6471 43.8318 73.1802 0 73.1802
5 1.0767 76.3061 Red Clay-Addisu Gebeya 2 15.3 21.3 42.5985 50.9499 91.4371 0 91.4371
6 1.0767 91.0843 Red Clay-Addisu Gebeya 2 15.3 21.3 47.5102 56.8246 107.298 0.793523 106.505
7 1.0767 104.94 Red Clay-Addisu Gebeya 2 15.3 21.3 51.1908 61.2268 120.278 2.48185 117.796
8 1.0767 117.867 Red Clay-Addisu Gebeya 2 15.3 21.3 54.0369 64.6308 130.284 3.75773 126.527
9 1.0767 127.259 Red Clay-Addisu Gebeya 2 15.3 21.3 55.2418 66.0719 134.839 4.61592 130.224
10 1.0767 124.365 Red Clay-Addisu Gebeya 2 15.3 21.3 52.299 62.5522 126.275 5.07905 121.196
11 1.0767 123.179 Red Clay-Addisu Gebeya 2 15.3 21.3 49.7188 59.4662 118.5 5.21899 113.281
12 1.0767 131.644 Red Clay-Addisu Gebeya 2 15.3 21.3 49.8462 59.6185 118.585 4.9138 113.672
13 1.0767 139.637 Red Clay-Addisu Gebeya 2 15.3 21.3 49.7697 59.5271 117.584 4.14768 113.437

MSc Thesis Page 90

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional Probabilistic Method in Slope Stability Analysis

14 1.0767 146.541 Red Clay-Addisu Gebeya 2 15.3 21.3 49.4531 59.1484 115.367 2.90119 112.466
15 1.0767 152.302 Red Clay-Addisu Gebeya 2 15.3 21.3 49.0114 58.6201 112.261 1.15059 111.11
16 1.0767 156.854 Red Clay-Addisu Gebeya 2 15.3 21.3 48.1665 57.6095 108.518 0 108.518
17 1.0767 160.116 Red Clay-Addisu Gebeya 2 15.3 21.3 46.8222 56.0017 104.394 0 104.394
18 1.0767 161.218 Red Clay-Addisu Gebeya 2 15.3 21.3 45.226 54.0925 99.4973 0 99.4973
19 1.0767 149.877 Red Clay-Addisu Gebeya 2 15.3 21.3 41.1465 49.2133 86.983 0 86.983
20 1.0767 133.33 Red Clay-Addisu Gebeya 2 15.3 21.3 44.5812 53.3213 97.5198 0 97.5198
21 1.0767 114.923 Red Clay-Addisu Gebeya 2 15.3 21.3 55.8559 66.8064 132.107 0 132.107
22 1.0767 94.3977 Red Clay-Addisu Gebeya 2 15.3 21.3 50.9706 60.9634 117.121 0 117.121
23 1.0767 71.4064 Red Clay-Addisu Gebeya 2 15.3 21.3 46.0008 55.0192 101.874 0 101.874
24 1.0767 45.4613 Red Clay-Addisu Gebeya 2 15.3 21.3 40.7414 48.7287 85.7398 0 85.7398
25 1.0767 15.8392 Red Clay-Addisu Gebeya 2 15.3 21.3 34.8698 41.706 67.7281 0 67.7281

MSc Thesis Page 91

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Interslice Data
• Global Minimum Query (bishop simplified) - Safety Factor: 1.20262

X Y Interslice Interslice Interslice

Slice
coordinate coordinate - Bottom Normal Force Shear Force Force Angle
Number
[m] [m] [kN] [kN] [degrees]

1 -0.108437 0 0 0 0
2 0.968268 -0.177082 18.9915 0 0
3 2.04497 -0.309011 45.7544 0 0
4 3.12168 -0.396465 78.2132 0 0
5 4.19838 -0.439882 114.48 0 0
6 5.27509 -0.439481 152.86 0 0
7 6.35179 -0.395259 191.55 0 0
8 7.4285 -0.306996 228.867 0 0
9 8.5052 -0.174246 263.699 0 0
10 9.58191 0.00367082 294.655 0 0
11 10.6586 0.227689 319.811 0 0
12 11.7353 0.49902 339.68 0 0
13 12.812 0.819193 354.768 0 0
14 13.8887 1.1901 364.515 0 0
15 14.9654 1.61404 368.438 0 0
16 16.0421 2.09384 366.17 0 0
17 17.1188 2.63293 357.055 0 0
18 18.1955 3.23552 340.332 0 0
19 19.2722 3.90679 315.984 0 0
20 20.349 4.65321 287.344 0 0
21 21.4257 5.48297 243.518 0 0
22 22.5024 6.40662 166.445 0 0
23 23.5791 7.43816 85.603 0 0
24 24.6558 8.59672 4.07186 0 0
25 25.7325 9.90963 -73.9257 0 0
26 26.8092 11.4184 0 0 0

MSc Thesis Page 92

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

• Global Minimum Query (janbu simplified) - Safety Factor: 1.05011

X Y Interslice Interslice Interslice
Slice
coordinate coordinate - Bottom Normal Force Shear Force Force Angle
Number
[m] [m] [kN] [kN] [degrees]
1 -1.76587 0 0 0 0
2 -0.628176 -0.50873 29.3824 0 0
3 0.509523 -0.938349 66.8728 0 0
4 1.64722 -1.29397 119.089 0 0
5 2.78492 -1.57949 181.853 0 0
6 3.92262 -1.79785 250.022 0 0
7 5.06032 -1.95118 320.507 0 0
8 6.19801 -2.04094 390.66 0 0
9 7.33571 -2.06793 458.183 0 0
10 8.47341 -2.03242 521.061 0 0
11 9.61111 -1.93408 576.734 0 0
12 10.7488 -1.772 621.239 0 0
13 11.8865 -1.54464 655.542 0 0
14 13.0242 -1.24977 680.816 0 0
15 14.1619 -0.884333 696.02 0 0
16 15.2996 -0.444279 700.239 0 0
17 16.4373 0.0756937 692.742 0 0
18 17.575 0.68252 673.014 0 0
19 18.7127 1.38535 640.811 0 0
20 19.8504 2.19649 597.77 0 0
21 20.9881 3.1329 549.025 0 0
22 22.1258 4.21894 450.038 0 0
23 23.2635 5.4916 335.843 0 0
24 24.4012 7.0119 214.994 0 0
25 25.5389 8.89491 96.3768 0 0
26 26.6766 11.4184 0 0 0

• Global Minimum Query (janbu corrected) - Safety Factor: 1.13014

X Y Interslice Interslice Interslice
Slice
coordinate coordinate - Bottom Normal Force Shear Force Force Angle
Number
[m] [m] [kN] [kN] [degrees]
1 -1.54488 0 0 0 0
2 -0.416024 -0.498757 28.8491 0 0
3 0.712835 -0.919653 66.9559 0 0
4 1.84169 -1.26764 120.311 0 0
5 2.97055 -1.54649 183.585 0 0
6 4.09941 -1.75904 251.883 0 0
7 5.22827 -1.90736 322.212 0 0
8 6.35713 -1.99284 391.999 0 0
9 7.48599 -2.01626 459.011 0 0
10 8.61485 -1.97782 521.289 0 0

MSc Thesis Page 93

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

11 9.7437 -1.87718 575.982 0 0

12 10.8726 -1.71342 619.503 0 0
13 12.0014 -1.48499 653.256 0 0
14 13.1303 -1.18966 678.062 0 0
15 14.2591 -0.824366 692.87 0 0
16 15.388 -0.385079 696.789 0 0
17 16.5169 0.133483 689.107 0 0
18 17.6457 0.73822 669.326 0 0
19 18.7746 1.43824 637.216 0 0
20 19.9034 2.24577 594.609 0 0
21 21.0323 3.17768 546.459 0 0
22 22.1612 4.2582 446.235 0 0
23 23.29 5.52411 332.98 0 0
24 24.4189 7.03612 213.172 0 0
25 25.5477 8.90869 95.5843 0 0
26 26.6766 11.4184 0 0 0

• Global Minimum Query (spencer) - Safety Factor: 1.19503

X Y Interslice Interslice Interslice
Slice
coordinate coordinate - Bottom Normal Force Shear Force Force Angle
Number
[m] [m] [kN] [kN] [degrees]
1 -0.329428 0 0 0 0
2 0.772027 -0.194664 23.6372 9.77223 22.4615
3 1.87348 -0.342453 57.0517 23.5866 22.4614
4 2.97494 -0.444141 97.4725 40.2976 22.4614
5 4.07639 -0.500253 142.175 58.7789 22.4615
6 5.17785 -0.511076 188.864 78.0811 22.4614
7 6.27931 -0.476664 235.059 97.1792 22.4614
8 7.38076 -0.396842 278.939 115.32 22.4614
9 8.48222 -0.271201 319.289 132.002 22.4614
10 9.58367 -0.0990887 354.636 146.616 22.4615
11 10.6851 0.120412 382.892 158.297 22.4614
12 11.7866 0.388503 404.885 167.39 22.4615
13 12.888 0.706711 421.411 174.222 22.4614
14 13.9895 1.07693 432.047 178.619 22.4614
15 15.091 1.50149 436.474 180.449 22.4614
16 16.1924 1.98323 434.514 179.639 22.4614
17 17.2939 2.52562 426.14 176.177 22.4614
18 18.3953 3.13293 410.54 169.728 22.4615
19 19.4968 3.8104 388.143 160.468 22.4614
20 20.5982 4.56458 362.486 149.861 22.4615
21 21.6997 5.40377 315.22 130.32 22.4615
22 22.8011 6.33868 248.567 102.764 22.4615
23 23.9026 7.38347 179.979 74.4077 22.4614
24 25.0041 8.55752 112.26 46.4113 22.4615

MSc Thesis Page 94

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

25 26.1055 9.88852 49.0007 20.2582 22.4615

26 27.207 11.4184 0 0 0

• Global Minimum Query (gle/morgenstern-price) - Safety Factor: 1.19605

X Y Interslice Interslice Interslice
Slice
coordinate coordinate - Bottom Normal Force Shear Force Force Angle
Number
[m] [m] [kN] [kN] [degrees]
1 -0.108437 0 0 0 0
2 0.968268 -0.177082 19.7909 1.36371 3.94179
3 2.04497 -0.309011 49.179 6.72399 7.78548
4 3.12168 -0.396465 86.2626 17.4585 11.4414
5 4.19838 -0.439882 128.852 34.1276 14.8347
6 5.27509 -0.439481 174.627 56.4314 17.9084
7 6.35179 -0.395259 220.977 83.165 20.6239
8 7.4285 -0.306996 265.414 112.433 22.9582
9 8.5052 -0.174246 306.232 142.152 24.9006
10 9.58191 0.00367082 341.652 169.957 26.4483
11 10.6586 0.227689 369.608 193.258 27.6038
12 11.7353 0.49902 390.926 211.117 28.371
13 12.812 0.819193 406.565 223.081 28.7534
14 13.8887 1.1901 416.477 228.519 28.7534
15 14.9654 1.61404 420.752 227.224 28.3709
16 16.0421 2.09384 419.599 219.397 27.6039
17 17.1188 2.63293 412.898 205.399 26.4484
18 18.1955 3.23552 400.346 185.839 24.9005
19 19.2722 3.90679 382.194 161.903 22.9582
20 20.349 4.65321 361.519 136.058 20.6239
21 21.4257 5.48297 328.546 106.171 17.9085
22 22.5024 6.40662 266.594 70.6098 14.8347
23 23.5791 7.43816 200.596 40.5981 11.4414
24 24.6558 8.59672 132.04 18.0532 7.7855
25 25.7325 9.90963 63.2856 4.36075 3.94179
26 26.8092 11.4184 0 0 0

Probabilistic Analysis Results (Global Minimum)

• Method: bishop simplified
• Factor of Safety, mean: 1.205084
• Factor of Safety, standard deviation: 0.087113
• Factor of Safety, minimum: 0.914353
• Factor of Safety, maximum: 1.526260
• Probability of Failure: 0.600% (= 6 failed surfaces / 1000 valid surfaces)
• Reliability index: 2.35422 (assuming normal distribution)
• Reliability index: 2.54790 (assuming lognormal distribution)
• Method: janbu simplified

MSc Thesis Page 95

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

• Factor of Safety, mean: 1.051945

• Factor of Safety, standard deviation: 0.073569
• Factor of Safety, minimum: 0.813378
• Factor of Safety, maximum: 1.304660
• Probability of Failure: 22.400% (= 224 failed surfaces / 1000 valid surfaces)
• Reliability index: 0.70607 (assuming normal distribution)
• Reliability index: 0.69005 (assuming lognormal distribution)
• Method: janbu corrected
• Factor of Safety, mean: 1.132124
• Factor of Safety, standard deviation: 0.079076
• Factor of Safety, minimum: 0.875657
• Factor of Safety, maximum: 1.403420
• Probability of Failure: 4.600% (= 46 failed surfaces / 1000 valid surfaces)
• Reliability index: 1.67084 (assuming normal distribution)
• Reliability index: 1.74393 (assuming lognormal distribution)
• Method: spencer
• Factor of Safety, mean: 1.201432
• Factor of Safety, standard deviation: 0.086172
• Factor of Safety, minimum: 0.915356
• Factor of Safety, maximum: 1.514600
• Probability of Failure: 0.600% (= 6 failed surfaces / 1000 valid surfaces)
• Reliability index: 2.33756 (assuming normal distribution)
• Reliability index: 2.52607 (assuming lognormal distribution)
• Method: gle/morgenstern-price
• Factor of Safety, mean: 1.201133
• Factor of Safety, standard deviation: 0.086621
• Factor of Safety, minimum: 0.912622
• Factor of Safety, maximum: 1.517440
• Probability of Failure: 0.600% (= 6 failed surfaces / 1000 valid surfaces)
• Reliability index: 2.32197 (assuming normal distribution)
• Reliability index: 2.50852 (assuming lognormal distribution)

List Of Coordinates

Water Table

X Y
-11.6 -2.31005
-2.8 -2.31164
9.648 0.549
19.637 2.987
28.098 5.935
33.715 8.807
36.904 10.2152

Line Load

MSc Thesis Page 96

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

X Y
28.0232 11.4184
21.0583 11.4184

Line Load

X Y
27.9815 11.4184
21.1417 11.4184

Line Load

X Y
-2.08851 0
-9.17853 0

Line Load

X Y
-2.0468 0
-9.26194 0

External Boundary

X Y
-11.6 -5.998
36.904 -5.998
36.904 14.171
33.11 14.171
30 11.4184
18.9293 11.4184
10.954 6.039
8.954 6.039
0 0
-11.6 0

MSc Thesis Page 97

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

APPENDIX B

Addisu gebeya 2.fez

RS2 Analysis Information

Project Summary
File Name: Addisu gebeya 2.fez
Last saved with RS2 version: 11.015
Analysis: Converted from Slide v6.02 with RS2 11.015

MSc Thesis Page 98

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

General Settings
Single stage model
Analysis Type: Plane Strain
Solver Type: Conjugate Gradient
Units: Metric, stress as kPa
Permeability Units: meters/second
Time Units: days

MSc Thesis Page 99

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Analysis Options
Maximum Number of Iterations: 500
Tolerance: 0.01
Number of Load Steps: Automatic
Convergence Type: Comprehensive
Tensile Failure: Reduces Shear Strength
Joint tension reduces joint stiffness by a factor of 0.01

MSc Thesis Page 100

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Strength Reduction Settings

Inital Estimate of SRF: 1
Step Size: Automatic
Tolerance (SRF): 0.01
Limit SSR Search Area: No
Accelerate SSR Analysis: No
Reduce SSR Iterations after failure: Yes
Apply SSR to Mohr-Coulomb Tensile
Yes
Strength:
Convergence Parameters: Automatic

MSc Thesis Page 101

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Groundwater Analysis
Method: Static
Pore Fluid Unit Weight: 9.81 kN/m3
Grid Interpolation: Modified Chugh
Probability: None

MSc Thesis Page 102

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Field Stress
Field stress: Gravity
Using actual ground surface
Effective stress ratio (horizontal/vertical in-
1
plane):
Effective stress ratio (horizontal/vertical out-
1
of-plane):
Locked-in horizontal stress (in-plane): 0
Locked-in horizontal stress (out-of-plane): 0

MSc Thesis Page 103

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Mesh
Mesh type: Uniform
Element type: 6 Noded triangles
Stage Name # of Elements # of Nodes
1 4929 10044
SRF: 1 4929 10044
SRF: 1.09 4929 10044
SRF: 1.14 4929 10044
SRF: 1.16 4929 10044
SRF: 1.17 4929 10044
SRF: 1.18 4929 10044
SRF: 1.2 4929 10044
SRF: 1.4 4929 10044

MSc Thesis Page 104

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Mesh Quality
All elements are of good quality

Poor quality elements defined as:

Side length ratio (maximum / minimum) > 30.00
Minimum interior angle < 2.0 degrees
Maximum interior angle > 175.0 degrees

MSc Thesis Page 105

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Excavation Areas
Original Un-deformed Areas
External Boundary Area: 619.622 m2
External Boundary Perimeter: 127.709 m

1
Values not available until this stage is viewed in a window

SRF: 1
Values not available until this stage is viewed in a window

SRF: 1.09
External Boundary Area: 619.660 m2 (0.0375843 m2 change from original area)
External Boundary Perimeter: 127.695 m (-0.0140382 m change from original
perimeter)

SRF: 1.14
External Boundary Area: 619.661 m2 (0.0391197 m2 change from original area)
External Boundary Perimeter: 127.698 m (-0.010532 m change from original
perimeter)

SRF: 1.16
External Boundary Area: 619.662 m2 (0.039709 m2 change from original area)
External Boundary Perimeter: 127.699 m (-0.00950395 m change from original
perimeter)

SRF: 1.17
External Boundary Area: 619.662 m2 (0.0400673 m2 change from original area)
External Boundary Perimeter: 127.700 m (-0.00884447 m change from original
perimeter)

SRF: 1.18
External Boundary Area: 619.664 m2 (0.0416655 m2 change from original area)
External Boundary Perimeter: 127.705 m (-0.00367586 m change from original
perimeter)

SRF: 1.2
External Boundary Area: 619.668 m2 (0.0457874 m2 change from original area)
External Boundary Perimeter: 127.719 m (0.0100214 m change from original
perimeter)

SRF: 1.4
External Boundary Area: 619.756 m2 (0.134321 m2 change from original area)
External Boundary Perimeter: 128.070 m (0.361119 m change from original
perimeter)

MSc Thesis Page 106

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Material Properties
Red Clay-Addisu Gebeya 2

Material Color
Initial Element Loading Field Stress and Body Force
Unit Weight 19.5 kN/m3
Elastic Type Isotropic
Poisson's Ratio 0.4
Young's Modulus 50000 kPa
Use Residual Young's Modulus No
Failure Criterion Mohr-Coulomb
Material Type Plastic
Peak Tensile Strength 39.24 kPa
Peak Friction Angle 21.3 degrees
Peak Cohesion 15.3 kPa
Residual Tensile Strength 39.24 kPa
Residual Friction Angle 21.3 degrees
Residual Cohesion 15.3 kPa
Dilation Angle 0 degrees
Apply SSR (Shear Strength Reduction) Yes
Use Unsaturated Parameters Yes
Unsaturated Shear Strength Angle 0 degrees
Air Entry Value 0
Material Behaviour Drained
Porosity Value 0.5
Static Water Mode Piezometric Lines
Piezo to Use 1

MSc Thesis Page 107

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Shear Strength Reduction - Material Properties

Strength Reduction Factor: 1
Maximum Total Displacement: 0.0221691 m
Converged: yes
Material Red Clay-Addisu Gebeya 2
Peak friction angle 21.3 degrees
Peak cohesion 15.3 kPa
Residual Friction Angle 21.3 degrees
Residual Cohesion 15.3 kPa

Strength Reduction Factor: 1.09

Maximum Total Displacement: 0.0227163 m
Converged: yes
Material Red Clay-Addisu Gebeya 2
Peak friction angle 19.6817 degrees
Peak cohesion 14.0367 kPa
Residual Friction Angle 19.6817 degrees
Residual Cohesion 14.0367 kPa

Strength Reduction Factor: 1.14

Maximum Total Displacement: 0.0230783 m
Converged: yes
Material Red Clay-Addisu Gebeya 2
Peak friction angle 18.8809 degrees
Peak cohesion 13.4211 kPa
Residual Friction Angle 18.8809 degrees
Residual Cohesion 13.4211 kPa

Strength Reduction Factor: 1.16

Maximum Total Displacement: 0.0232529 m
Converged: yes
Material Red Clay-Addisu Gebeya 2
Peak friction angle 18.5778 degrees
Peak cohesion 13.1897 kPa
Residual Friction Angle 18.5778 degrees
Residual Cohesion 13.1897 kPa

Critical Strength Reduction Factor: 1.17

Maximum Total Displacement: 0.0233477 m
Converged: yes
Material Red Clay-Addisu Gebeya 2
Peak friction angle 18.4298 degrees
Peak cohesion 13.0769 kPa
Residual Friction Angle 18.4298 degrees
Residual Cohesion 13.0769 kPa

MSc Thesis Page 108

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Strength Reduction Factor: 1.18

Maximum Total Displacement: 0.0234439 m

Converged: no
Material Red Clay-Addisu Gebeya 2
Peak friction angle 18.2841 degrees
Peak cohesion 12.9661 kPa
Residual Friction Angle 18.2841 degrees
Residual Cohesion 12.9661 kPa

Strength Reduction Factor: 1.2

Maximum Total Displacement: 0.0369142 m
Converged: no
Material Red Clay-Addisu Gebeya 2
Peak friction angle 17.9991 degrees
Peak cohesion 12.75 kPa
Residual Friction Angle 17.9991 degrees
Residual Cohesion 12.75 kPa

Strength Reduction Factor: 1.4

Maximum Total Displacement: 0.326274 m
Converged: no
Material Red Clay-Addisu Gebeya 2
Peak friction angle 15.5619 degrees
Peak cohesion 10.9286 kPa
Residual Friction Angle 15.5619 degrees
Residual Cohesion 10.9286 kPa

MSc Thesis Page 109

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Displacements
Displacement data is not available until total displacement is viewed in a window

Yielded Elements
Yielded Mesh Elements
Number of yielded mesh elements is not available for 1 until the stage is viewed in
a window
Number of yielded mesh elements is not available for SRF: 1 until the stage is
viewed in a window
Number of yielded mesh elements on
642
SRF: 1.09:
Number of yielded mesh elements on
713
SRF: 1.14:
Number of yielded mesh elements on
751
SRF: 1.16:
Number of yielded mesh elements on
774
SRF: 1.17:
Number of yielded mesh elements on
795
SRF: 1.18:
Number of yielded mesh elements on
839
SRF: 1.2:
Number of yielded mesh elements on
1602
SRF: 1.4:

MSc Thesis Page 110

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

List of All Coordinates

External boundary
X Y
-11.6 -5.998
36.904 -5.998
36.904 14.171
33.11 14.171
30 11.4184
28.0462 11.4184
28.0212 11.4184
21.0048 11.4184
18.9293 11.4184
10.954 6.039
8.954 6.039
0 0
-0.997124 0
-7.9 0
-11.6 0

Piezometric line
X Y
-11.6 -2.31005
-2.8 -2.31164
9.648 0.549
19.637 2.987
28.098 5.935
33.715 8.807
36.904 10.2152

MSc Thesis Page 111

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

APPENDIX C

PSSA Addisu gebeya 2.fez

RS2 Analysis Information

Project Summary
File Name: PSSA Addisu gebeya 2.fez
Last saved with RS2 version: 11.015
Analysis: Converted from Slide v6.02 with RS2 11.015

General Settings
Single stage model
Analysis Type: Plane Strain
Solver Type: Conjugate Gradient
Units: Metric, stress as kPa
Permeability Units: meters/second
Time Units: days

Options
Maximum Number of Iterations: 500
Tolerance: 0.01
Number of Load Steps: Automatic
Convergence Type: Comprehensive
Tensile Failure: Reduces Shear Strength
Joint tension reduces joint stiffness by a factor of 0.01

Strength Reduction Settings

Inital Estimate of SRF: 1
Step Size: Automatic
Tolerance (SRF): 0.01
Limit SSR Search Area: No
Accelerate SSR Analysis: No
Reduce SSR Iterations after failure: Yes
Apply SSR to Mohr-Coulomb Tensile
Yes
Strength:
Convergence Parameters: Automatic

MSc Thesis Page 112

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Groundwater Analysis
Method: Static
Pore Fluid Unit Weight: 9.81 kN/m3
Grid Interpolation: Modified Chugh

Probability Analysis
Analysis Type: Latin-Hypercube
Number of Samples: 100

Statistical Variables
Number of Statistical Variables Used: 2
Variable Material/Joint Property Standard
Distribution Mean Min Max
Type Name Type Deviation
Friction
Material Red Clay-Addisu
Angle Normal 21.3 1.704 16.188 26.412
Property Gebeya 2
(peak)
Material Red Clay-Addisu Cohesion
Normal 15.3 3.825 3.825 26.775
Property Gebeya 2 (peak)

Field Stress
Field stress: Gravity
Using actual ground surface
Effective stress ratio (horizontal/vertical in-
1
plane):
Effective stress ratio (horizontal/vertical out-
1
of-plane):
Locked-in horizontal stress (in-plane): 0
Locked-in horizontal stress (out-of-plane): 0

MSc Thesis Page 113

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Mesh
Mesh type: Uniform
Element type: 6 Noded triangles
Stage Name # of Elements # of Nodes
1 4929 10044
SRF: 0.49 4929 10044
SRF: 0.74 4929 10044
SRF: 0.86 4929 10044
SRF: 0.92 4929 10044
SRF: 0.95 4929 10044
SRF: 0.97 4929 10044
SRF: 0.98 4929 10044
SRF: 0.99 4929 10044
SRF: 1 4929 10044

Mesh Quality
All elements are of good quality

Poor quality elements defined as:

Side length ratio (maximum / minimum) > 30.00
Minimum interior angle < 2.0 degrees
Maximum interior angle > 175.0 degrees

Excavation Areas
Original Un-deformed Areas
External Boundary Area: 619.622 m2
External Boundary Perimeter: 127.709 m

1
External Boundary Area: 619.620 m2 (-0.00210763 m2 change from original area)
External Boundary Perimeter: 127.680 m (-0.0292687 m change from original
perimeter)

SRF: 0.49
Values not available until this stage is viewed in a window

SRF: 0.74
Values not available until this stage is viewed in a window

MSc Thesis Page 114

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

SRF: 0.86
Values not available until this stage is viewed in a window

SRF: 0.92
Values not available until this stage is viewed in a window

SRF: 0.95
Values not available until this stage is viewed in a window

SRF: 0.97
External Boundary Area: 619.620 m2 (-0.00164973 m2 change from original area)
External Boundary Perimeter: 127.680 m (-0.0290156 m change from original
perimeter)

SRF: 0.98
Values not available until this stage is viewed in a window

SRF: 0.99
Values not available until this stage is viewed in a window

SRF: 1
Values not available until this stage is viewed in a window

Material Properties
Red Clay-Addisu Gebeya 2

Material Color
Initial Element Loading Field Stress and Body Force
Unit Weight 19.5 kN/m3
Elastic Type Isotropic
Poisson's Ratio 0.4
Young's Modulus 50000 kPa
Use Residual Young's Modulus No
Failure Criterion Mohr-Coulomb
Material Type Plastic
Peak Tensile Strength 39.24 kPa
Peak Friction Angle 21.3 degrees
Peak Cohesion 15.3 kPa
Residual Tensile Strength 39.24 kPa
Residual Friction Angle 21.3 degrees
Residual Cohesion 15.3 kPa
Dilation Angle 0 degrees
Apply SSR (Shear Strength Reduction) Yes
Use Unsaturated Parameters Yes
Unsaturated Shear Strength Angle 0 degrees

MSc Thesis Page 115

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Air Entry Value 0

Material Behaviour Undrained
Fluid Bulk Modulus 2.2e+06 kPa
Porosity Value 0.3
Static Water Mode Piezometric Lines
Piezo to Use 1

Strength Reduction Factor Statistics

Probability of Failure: 100 %
Critical SRF mean value: 0.97
Critical SRF std. dev. value: 0.0024
Critical SRF is present only in 69 out of
NOTE:
100 files
File # Critical SRF
1 0.97
2 0.97
3 0.97
4 0.97
5 0.96
6 0.97
7 0.97
8 0.97
9 0.98
10 0.97
11 0.97
12 0.97
13 0.96
14 0.97
15 0.97
16 0.97
17 0.97
18 0.97
19 0.97
20 0.97
21 0.97
22 0.97
23 0.97
24 0.97
25 0.97
26 0.97
27 0.97
28 0.97
29 0.97
30 0.97
31 0.97
32 0.97
33 0.97
34 0.97
35 0.97
36 0.97

MSc Thesis Page 116

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

37 0.97
38 0.97
39 0.97
40 0.97
41 0.97
42 0.97
43 0.97
44 0.97
45 0.97
46 0.97
47 0.97
48 0.97
49 0.97
50 0.97
51 0.97
52 0.97
53 0.97
54 0.97
55 0.97
56 0.97
57 0.97
58 0.97
59 0.97
60 0.97
61 0.97
62 0.97
63 0.97
64 0.98
65 0.97
66 0.97
67 0.97
68 0.97
69 0.97
70 None Converged
71 None Converged
72 None Converged
73 None Converged
74 None Converged
75 None Converged
76 None Converged
77 None Converged
78 None Converged
79 None Converged
80 None Converged
81 None Converged
82 None Converged
83 None Converged
84 None Converged
85 None Converged
86 None Converged
87 None Converged
88 None Converged
89 None Converged
90 None Converged
91 None Converged
92 None Converged

MSc Thesis Page 117

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

93 None Converged
94 None Converged
95 None Converged
96 None Converged
97 None Converged
98 None Converged
99 None Converged
100 None Converged

Probability of Failure: 100 %

Critical SRF mean value: 0.97
Critical SRF std. dev. value: 0.0024
Critical SRF is present only in 69 out of
NOTE:
100 files
File # Critical SRF
1 0.97
2 0.97
3 0.97
4 0.97
5 0.96
6 0.97
7 0.97
8 0.97
9 0.98
10 0.97
11 0.97
12 0.97
13 0.96
14 0.97
15 0.97
16 0.97
17 0.97
18 0.97
19 0.97
20 0.97
21 0.97
22 0.97
23 0.97
24 0.97
25 0.97
26 0.97
27 0.97
28 0.97
29 0.97
30 0.97
31 0.97
32 0.97
33 0.97
34 0.97
35 0.97
36 0.97

MSc Thesis Page 118

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

37 0.97
38 0.97
39 0.97
40 0.97
41 0.97
42 0.97
43 0.97
44 0.97
45 0.97
46 0.97
47 0.97
48 0.97
49 0.97
50 0.97
51 0.97
52 0.97
53 0.97
54 0.97
55 0.97
56 0.97
57 0.97
58 0.97
59 0.97
60 0.97
61 0.97
62 0.97
63 0.97
64 0.98
65 0.97
66 0.97
67 0.97
68 0.97
69 0.97
70 None Converged
71 None Converged
72 None Converged
73 None Converged
74 None Converged
75 None Converged
76 None Converged
77 None Converged
78 None Converged
79 None Converged
80 None Converged
81 None Converged
82 None Converged
83 None Converged
84 None Converged
85 None Converged
86 None Converged
87 None Converged
88 None Converged
89 None Converged
90 None Converged
91 None Converged
92 None Converged

MSc Thesis Page 119

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

93 None Converged
94 None Converged
95 None Converged
96 None Converged
97 None Converged
98 None Converged
99 None Converged
100 None Converged

Probability of Failure: 100 %

Critical SRF mean value: 0.97
Critical SRF std. dev. value: 0.0024
Critical SRF is present only in 69 out of
NOTE:
100 files
File # Critical SRF
1 0.97
2 0.97
3 0.97
4 0.97
5 0.96
6 0.97
7 0.97
8 0.97
9 0.98
10 0.97
11 0.97
12 0.97
13 0.96
14 0.97
15 0.97
16 0.97
17 0.97
18 0.97
19 0.97
20 0.97
21 0.97
22 0.97
23 0.97
24 0.97
25 0.97
26 0.97
27 0.97
28 0.97
29 0.97
30 0.97
31 0.97
32 0.97
33 0.97

MSc Thesis Page 120

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

34 0.97
35 0.97
36 0.97
37 0.97
38 0.97
39 0.97
40 0.97
41 0.97
42 0.97
43 0.97
44 0.97
45 0.97
46 0.97
47 0.97
48 0.97
49 0.97
50 0.97
51 0.97
52 0.97
53 0.97
54 0.97
55 0.97
56 0.97
57 0.97
58 0.97
59 0.97
60 0.97
61 0.97
62 0.97
63 0.97
64 0.98
65 0.97
66 0.97
67 0.97
68 0.97
69 0.97
70 None Converged
71 None Converged
72 None Converged
73 None Converged
74 None Converged
75 None Converged
76 None Converged
77 None Converged
78 None Converged
79 None Converged
80 None Converged
81 None Converged
82 None Converged
83 None Converged
84 None Converged
85 None Converged
86 None Converged
87 None Converged
88 None Converged
89 None Converged

MSc Thesis Page 121

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

90 None Converged
91 None Converged
92 None Converged
93 None Converged
94 None Converged
95 None Converged
96 None Converged
97 None Converged
98 None Converged
99 None Converged
100 None Converged

Shear Strength Reduction - Material Properties

Strength Reduction Factor: 0.49
Maximum Total Displacement: 0.0184936 m
Converged: yes
Material Red Clay-Addisu Gebeya 2
Peak friction angle 38.5086 degrees
Peak cohesion 31.2245 kPa
Residual Friction Angle 38.5086 degrees
Residual Cohesion 31.2245 kPa

Strength Reduction Factor: 0.74

Maximum Total Displacement: 0.0192849 m
Converged: yes
Material Red Clay-Addisu Gebeya 2
Peak friction angle 27.7834 degrees
Peak cohesion 20.6757 kPa
Residual Friction Angle 27.7834 degrees
Residual Cohesion 20.6757 kPa

Strength Reduction Factor: 0.86

Maximum Total Displacement: 0.019812 m
Converged: yes
Material Red Clay-Addisu Gebeya 2
Peak friction angle 24.3873 degrees
Peak cohesion 17.7907 kPa
Residual Friction Angle 24.3873 degrees
Residual Cohesion 17.7907 kPa

Strength Reduction Factor: 0.92

Maximum Total Displacement: 0.0201119 m
Converged: yes
Material Red Clay-Addisu Gebeya 2
Peak friction angle 22.9666 degrees

MSc Thesis Page 122

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Peak cohesion 16.6304 kPa

Residual Friction Angle 22.9666 degrees
Residual Cohesion 16.6304 kPa

Strength Reduction Factor: 0.95

Maximum Total Displacement: 0.0202708 m
Converged: yes
Material Red Clay-Addisu Gebeya 2
Peak friction angle 22.3134 degrees
Peak cohesion 16.1053 kPa
Residual Friction Angle 22.3134 degrees
Residual Cohesion 16.1053 kPa

Critical Strength Reduction Factor: 0.97

Maximum Total Displacement: 0.0203829 m
Converged: yes
Material Red Clay-Addisu Gebeya 2
Peak friction angle 21.8973 degrees
Peak cohesion 15.7732 kPa
Residual Friction Angle 21.8973 degrees
Residual Cohesion 15.7732 kPa

Strength Reduction Factor: 0.98

Maximum Total Displacement: 0.0204394 m
Converged: no
Material Red Clay-Addisu Gebeya 2
Peak friction angle 21.6947 degrees
Peak cohesion 15.6122 kPa
Residual Friction Angle 21.6947 degrees
Residual Cohesion 15.6122 kPa

Strength Reduction Factor: 0.99

Maximum Total Displacement: 0.020494 m
Converged: no
Material Red Clay-Addisu Gebeya 2
Peak friction angle 21.4956 degrees
Peak cohesion 15.4545 kPa
Residual Friction Angle 21.4956 degrees
Residual Cohesion 15.4545 kPa

Strength Reduction Factor: 1

Maximum Total Displacement: 0.0206172 m
Converged: no
Material Red Clay-Addisu Gebeya 2
Peak friction angle 21.3 degrees
Peak cohesion 15.3 kPa
Residual Friction Angle 21.3 degrees
Residual Cohesion 15.3 kPa

Displacements

MSc Thesis Page 123

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Displacement data is not available until total displacement is viewed in a window

Yielded Elements
Yielded Mesh Elements
Number of yielded mesh elements on 1: 542
Number of yielded mesh elements is not available for SRF: 0.49 until the stage is
viewed in a window
Number of yielded mesh elements is not available for SRF: 0.74 until the stage is
viewed in a window
Number of yielded mesh elements is not available for SRF: 0.86 until the stage is
viewed in a window
Number of yielded mesh elements is not available for SRF: 0.92 until the stage is
viewed in a window
Number of yielded mesh elements is not available for SRF: 0.95 until the stage is
viewed in a window
Number of yielded mesh elements on
503
SRF: 0.97:
Number of yielded mesh elements is not available for SRF: 0.98 until the stage is
viewed in a window
Number of yielded mesh elements is not available for SRF: 0.99 until the stage is
viewed in a window
Number of yielded mesh elements is not available for SRF: 1 until the stage is
viewed in a window

List of All Coordinates

External boundary
X Y
-11.6 -5.998
36.904 -5.998
36.904 14.171
33.11 14.171
30 11.4184
28.0462 11.4184
28.0212 11.4184
21.0048 11.4184
18.9293 11.4184
10.954 6.039
8.954 6.039
0 0
-0.997124 0
-7.9 0
-11.6 0

MSc Thesis Page 124

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Piezometric line
X Y
-11.6 -2.31005
-2.8 -2.31164
9.648 0.549
19.637 2.987
28.098 5.935
33.715 8.807
36.904 10.2152

MSc Thesis Page 125

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

APPENDIX D

Minimum Slip Surfaces

SLIDE V6.0

Kolfe 1

FS (deterministic) = 1.194
FS overall(mean) = 1.197
PF = 5.400%
RI (normal) = 1.580
RI (lognormal) = 1.679

Kolfe 2

FS (deterministic) = 1.355
FS overall(mean) = 1.359
PF = 0.200%
RI (normal) = 3.148
RI (lognormal) = 3.619

Kolfe 3

MSc Thesis Page 126

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

FS (deterministic) = 2.872
FS overall(mean) = 2.884
PF = 0.300%
RI (normal) = 2.544
RI (lognormal) = 4.066

Addisu gebeya 1

FS (deterministic) = 1.341
FS overall(mean) = 1.347
PF = 2.300%
RI (normal) = 2.063
RI (lognormal) = 2.333

Addisu gebeya 2

FS (deterministic) = 1.203
FS overall(mean) = 1.205
PF = 0.600%
RI (normal) = 2.354
RI (lognormal) = 2.548

MSc Thesis Page 127

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Arada

FS (deterministic) = 2.265
FS overall(mean) = 2.272
PF = 0.000%
RI (normal) = 3.164
RI (lognormal) = 4.587

Asko

FS (deterministic) = 1.411
FS overall(mean) = 1.414
PF = 0.300%
RI (normal) = 2.766
RI (lognormal) = 3.229

RS2 V11

Kolfe 1

MSc Thesis Page 128

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Kolfe 2

Kolfe 3

Addisu gebeya 1

Addisu gebeya 2

MSc Thesis Page 129

 Reliability Of Probabilistic Finite Element Method Over Deterministic and Traditional
Probabilistic Method in Slope Stability Analysis

Arada

Asko

MSc Thesis Page 130