MATEC Web of Conferences 394, 01005 (2024) https://doi.org/10.

1051/matecconf/202439401005
ICCEE2023

A comparative study of critical failure surface in unsaturated

soil slopes.
Abderrahim Mihoubi 1*, Abdelkader Houam 1, Samir Benmoussa2, Mohamed Salah Laouar1, Ali
Fernane1, Salah Messast 3 and Saida Gueraidia 1
1
Mining Laboratory, Department of Civil Engineering, 12002 University of Tébessa, Algeria.
2
LGC-ROI, Department of Civil Engineering University of Batna 2, 05078 Batna, Algeria.
3
Department of Civil Engineering, 21000 University of Skikda, Algeria.

Abstract. Landslides are rapid downslope movements triggered by factors such as rainfall, earthquakes, or
human activity, where factors like relative permeability, air flow, and variations in hydraulic conductivity
affect the mechanisms behind rainfall infiltration and landslide occurrence. This study investigates these
phenomena using the finite difference method (FDM) and finite element method (FEM) in FLAC and
PLAXIS software. These methods account for the coexistence of water and air, including the Two-phase
flow option, in evaluating unsaturated slope stability and identifying critical failure surfaces. The efficiency
of these approaches is gauged by varying soil parameters and exploring the impact of different values of
suction and hydraulic conductivity, as well as soil-water characteristics, on the safety factor. It was observed
that there are disparities between both software packages at high suction values, which impact the safety
factor. FLAC is noted for its conservative approach and superiority in representing unfavourable conditions.

1 Introduction investigations have delved into the correlation between

slope stability and rainfall intensity and duration. The
Predicting the risk of landslides due to rainfall is crucial, earlier findings have underscored that analysing slope
especially in mountainous regions and steep areas. stability in an unsaturated state is a more practical
These landslides occur when the soil loses its cohesion approach, as it accounts for the actual groundwater level
and moves downhill, posing a threat to infrastructure (GWL) position and the alterations in shear strength due
and human lives. Various factors contribute to increased to deformation and fluctuations in pore-water pressure
landslide risks, including soil characteristics, geological [5].
relief, precipitation intensity, terrain slope, and human With advancements in technology, the utilization of
activities, where the slope failures triggered by rainfall software packages employing finite difference and finite
have frequently been reported worldwide. element methods has gained significant popularity.
There are two primary techniques for evaluating how Currently, slope stability poses a prominent challenge
unsaturated soils react to rainfall infiltration: the for soil professionals and engineers, with ongoing
coupled and uncoupled methods. In the uncoupled debate among scientists regarding the most effective
approach, the first step involves conducting seepage approach for slope stability assessment.
analysis to predict pore water pressures. These pressures This study's primary objectives encompass the
are then utilized as data for deformation or stability utilization of various slope stability analysis software
analysis. On the other hand, the coupled approach takes tools to assess the influence of site-specific hydrological
into account the mutual influence between seepage and characteristics, specifically the Soil-Water
deformation by simultaneously solving for pore water Characteristic Curve (SWCC) and the rainfall intensity,
pressures and deformation during rainfall infiltration. on the safety and the determination of critical slip
This integrated approach provides a more realistic surfaces. Furthermore, this research aims to compare the
representation of the physical interaction between water outcomes derived from different slope stability analysis
flow and stress-strain changes within unsaturated soils. software programs. To achieve these goals, three
Previous research has made substantial contributions distinct soil types were obtained from the Tebessa
to the field of unsaturated slope stability analysis. These region, each characterized by unique hydrological
studies have revealed that the coupled hydro-mechanical parameters, are simulated and analysed within this
model (CHM) yields the most accurate results when study.
simulating rainfall infiltration [10], Additionally, these

* Corresponding author: [email protected]

© The Authors, published by EDP Sciences. This is an open access article distributed under the terms of the Creative Commons Attribution License 4.0
(https://creativecommons.org/licenses/by/4.0/).
MATEC Web of Conferences 394, 01005 (2024) https://doi.org/10.1051/matecconf/202439401005
ICCEE2023

2 Model formulation A commonly employed approach to represent the

hydraulic properties of unsaturated soils is the set of
closed-form equations introduced by van Genuchten.
2.1 Slope stability analysis by finite elements These equations are built upon Mualem's capillary
and finite difference model. The capillary pressure law used in FLAC is in
The finite element analysis (FEA) and finite difference the following van Genuchten form:
1 m

Pc  P0  S e  m  1
(FDM) locates the critical failure surface based on a 
1

(3)
calculation of failure shear strain zones which have been  
shown to coincide with rupture surfaces in experimental
tests. The method of calculation of FoS in PLAXIS 2D w g
P0  (4)
and FLAC 2D is the shear strength reduction method. 
When analysing unsaturated slopes in FLAC, Where (Se) is effective saturation; m and P0 are the
suction is typically incorporated through constitutive shape parameters and P0 is the reference capillary
models such as the Soil Water Characteristic Curve pressure.
(SWCC) and the unsaturated soil shear strength Equation (5) defines the relationship between the
parameters. The impact of suction on the safety factor of mobility coefficient utilized in FLAC, which
unsaturated slopes in FLAC is influenced by the explicit characterizes permeability, and the hydraulic
representation of soil-water-air interactions, allowing conductivity commonly employed in expressing Darcy's
for a detailed analysis of moisture-induced changes in law in terms of head.
soil behaviour and its effect on slope stability. On the Ks
other hand, PLAXIS utilizes a finite element method K (5)
( g w )
(FEM) for numerical analysis, focusing on the
discretization of the soil domain into finite elements.
PLAXIS offers various constitutive models to account 2.3 Slope Parameters
for unsaturated soil behaviour, including the Van
Genuchten model for the soil-water retention curve. In The primary aim of this comparative study is to assess
PLAXIS, suction influences the safety factor of soil infiltration processes using FLAC and PLAXIS
unsaturated slopes by affecting soil shear strength software while exploring how the hydrological and
parameters and pore-water pressure distributions. mechanical properties of soils impact the slope stability
To numerically model the initiation of slope failures and length of the arc of failure. The slope angle of
caused by rainfall under unsaturated conditions, it is approximately 34 degrees was selected, as shown (see
necessary to develop a hydromechanically coupled Fig. 1). which was considered for the analysis.
transient analysis in a systematic way. In the case of
unsaturated soil, the pore volume is partially filled with
water, while the remaining portion is occupied by air.
Consequently, the key disparity between the shear
strength of saturated and unsaturated soils lies in the
definition of effective normal stress.
A practical Equation (1) provided by Bishop is used
in PLAXIS and FLAC to define the shear strength of
unsaturated soil. This equation incorporates parameters
like the frictional resistance angle (  ), pore air pressure
(Pa), and pore water pressure (Pw), and it is expressed
as:
Fig .1. Slope geometry and boundary conditions adopted in
max  (  Pa ) tan   ( Pa  Pw ) tan   c this study.

the safety factor is obtained as: Field and laboratory tests, direct shear tests, and
soil-water characteristic curve (SWCC), were conducted
 max to obtain the necessary material properties of the soil.
Fs  (2)
n These tests provide crucial data for accurately
simulating the behaviour of the slope and analysing the
Where (τn) is the reduced shear strength that is just large infiltration process. When constructing the model, the
enough to maintain equilibrium. layer inputs were tailored to the specific properties of
the soil slope being simulated. A detailed description of
2.2 Unsaturated flow these inputs can be found in the provided (see Fig. 2 and
Table.1).
Matric suction is described as the difference between
pore air pressure and pore-water pressure. The
connection between soil water content and pore-water
pressure is represented as the soil-water characteristic
curve (SWCC), which plots volumetric moisture content
against matric suction.

2
MATEC Web of Conferences 394, 01005 (2024) https://doi.org/10.1051/matecconf/202439401005
ICCEE2023

provide a clear illustration of the progressive seepage

phenomena resulting from the infiltration of rainfall.

Fig .2. SWCCs used in this study.

Table 1. Properties of soil water, and air.

Parameter Soil 1 Soil 2 Soil 3

Dry density γd (kN/m3) 18 17.5 20
3 3
Bulk modulus K (kPa) 9.8 × 10 6.67 × 10 2.33 ×104

Shear modulus G (kPa) 3.76 × 103 3.07 × 103 5 × 103

Cohesion C (kPa) 12 15 31

Friction angle Φ 21 19 13.5

Hydraulic conductivity 1.81 ×10-6 3.94 ×10-6 1.42 ×10-6
Ks (m/sec)
Shape parameter α 0.247 0.067 0.09
Shape parameter m 0.5306 0.5302 0.547
Reference capillary 39.77 146.42 108.95
pressure P0 (kPa)

3 Results and discussion

3.1 Comparison of suction profile

When assessing the factor of safety in unsaturated
slopes, it's crucial to examine the suction profile, which
represents hydrostatic conditions after rainfall
infiltration. Consequently, a predictive study was
performed for each soil type as shown in (Figure 3), Fig .3. Vertical profile of matric suction in various soil type.
subjecting them to a 10 mm/h rainfall intensity over a 5-
day duration. It's noteworthy that the comparison was Following five days of continuous rainfall, a notable
conducted rigorously, considering the initial suction disparity emerges in the behaviour of the PLAXIS 2D
values, which were set at 50% of initial saturation in all pore pressure profile when compared to the FLAC
cases. Multiple simulation tests were carried out to software. It is worth noting that both software programs
confirm the accuracy of the approach using previous yield nearly identical pressure values at depths of 1
studies [4]. meter below the surface and 10 meters, but intriguingly,
In the context of seepage analysis, the initial suction the disparities in pressure values become most apparent
values for distinct soil types hold considerable within the depth range of 2 to 9 meters. This discrepancy
significance. The first soil type is characterized by a can be attributed to variations in the permeability
steady-state suction of -69 kPa at 50 % of initial equations employed by the two software platforms.
saturation, whereas the second type exhibits a In general, the use of higher-nodded elements,
significantly lower value of -200 kPa, and the third soil whether in triangular, quadrilateral, or mixed
type is associated with a suction level of -141 kPa. These configurations, offers a more robust and accurate
distinctive suction profiles, as showcased in Figure 3, solution due to the increased number of integrations

3
MATEC Web of Conferences 394, 01005 (2024) https://doi.org/10.1051/matecconf/202439401005
ICCEE2023

points during the numerical integration process in finite 3.3 Impact of rainfall intensity on the slip
element seepage analysis. surface length
In this study, the impact of rainfall intensity on the slip
3.2 Safety factor comparison surface was carefully examined, with a particular focus
on the length of the failure arc (La) as a quantitative
To inspect the outputs from PLAXIS and FLAC the
variable. To accurately assess this, precise
models were chosen up to their properties and exposed
measurements were obtained by creating circles using
to various rainfall intensities as shown in Tables 2 which
AutoCAD software to identify entry and exit points
summarizes the distinction between the FOS acquired
within the slope area, allowing for a comprehensive
from both programs and compares these results using
analysis of the failure arc.
the following formula:

FOS(FLAC) − FOS(PLAXIS)
𝐷𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑐𝑒 = × 100
FOS(FLAC) + FOS(PLAXIS)
2
Table 2. The difference of FOS between FLAC and
PLAXIS.
Rain FLAC PLAXIS Difference
Soil
Intensity FOS FOS %
Type
(mm/h)
0 1.64 1.662 -1.33
2.5 1.611 1.637 -1.6
1 5 1.576 1.534 -2.7
7.5 1.533 1.274 18.45
10 1.463 1.175 21.8
0 2.225 3.25 -37.44 Fig .4. Failure arc creation in AutoCAD
2.5 2.209 3.196 -36.52
2 5 2.186 3.973 -30.5 Inclusion of Figures (5) in this study offer a
7.5 2.158 2.782 -25.26 comprehensive visual depiction of the research findings.
10 2.127 2.415 -12.68 These figures elucidate the relationship between
0 1.762 1.926 -8.89 hydrological characteristics and rainfall intensity,
2.5 1.74 1.901 -8.84 shedding light on their collective impact on the variation
3 5 1.725 1.862 -7.63 in the length of the failure arc within the unsaturated
7.5 1.706 1.761 -3.17 slope systems under investigation. Through meticulous
10 1.678 1.663 0.89 analysis, it becomes evident that heightened rainfall
intensity correlates directly with a discernible reduction
The influence of rainfall intensity and hydrologic in the length of the surface failure arc.
characteristics on the safety factor is visually Moreover, an observation arises from the
represented in the table. Notably, there is a significant comparative analysis of the results obtained from FLAC
difference in FOS values between the FLAC and and PLAXIS. Despite FLAC exhibiting lower safety
PLAXIS programs in certain cases. For the first type of factor values, it is intriguing to note that the length of
soil, under steady-state conditions and with rainfall the failure surface in FLAC surpasses that in PLAXIS.
intensity less than 5 mm, there was no difference. This observation unveils a compelling insight into the
However, it is important to note that the greater the nuanced interplay between numerical modelling
initial suction, the larger the difference in the safety methodologies and the resulting surface failure
factor. As rainfall rate increased, the disparity also characteristics.
increased. It is noticeable that FOS values obtained from
FLAC are lower in 80% of the scenarios, and the rate of
decrease in FOS from 0 to 10 mm/h in PLAXIS is
greater than in FLAC.

Fig .5. The effect of IR on La in different types of soils.

4
MATEC Web of Conferences 394, 01005 (2024) https://doi.org/10.1051/matecconf/202439401005
ICCEE2023

It suggests that the choice of software platform can exert 5. H-E. Chen, T-L, Tsai, J-C, Yang, Threshold of
a discernible impact on the morphology of failure Slope Instability Induced by Rainfall and Lateral
surfaces within unsaturated slope systems. Such Flow. Water. 9, 722 (2017)
observations prompt a deeper exploration into the https://doi.org/10.3390/w9090722
underlying mechanisms driving these variations, 6. D.G. Fredlund, N.R. Morgenstern, R.A. Widger,
necessitating further investigation and analysis to The shear strength of unsaturated soils. Canadian
elucidate the factors influencing the discrepancy Geotechnical Journal. 15, 313–321 (1978)
between FLAC and PLAXIS results in terms of failure
https://doi.org/10.1139/t78-029
surface length. This insight underscores the importance
of meticulous consideration when selecting a numerical 7. S.L. Barbour, D.G. Fredlund, M.D. Fredlund, A.
modelling approach, as it can profoundly shape the Xing, The relationship of the unsaturated soil shear
outcomes and interpretations derived from slope strength to the soil-water characteristic curve.
stability analyses. Canadian Geotechnical Journal. 33, 440–448
(1995)
https://doi.org/10.1139/t96-065
4 Conclusions
8. S.G. Goh, E.C. Leong, H.Rahardjo, Shear Strength
This comparative review of popular computer programs Equations for Unsaturated Soil under Drying and
for rainfall-induced seepage unsaturated slope problems Wetting. Journal of Geotechnical and Geo-
highlights the direct relationship between hydrological environmental Engineering. 136, 594-606 (2010)
characteristics, rainfall intensities, and safety factors. https://doi.org/10.1061/(ASCE)GT.1943-
Notably, in fully coupled numerical models using FLAC 5606.0000261
and PLAXIS 2D, hydraulic conductivity (ks) emerges as 9. Itasca Manual. Fluid–Mechanical Interaction,
the primary controlling parameter for the FLAC 8.0, Itasca Consulting Group, Inc (2011)
hydromechanical response and failure time of
unsaturated slopes subjected to rainfall infiltration. 10. S.U. Cho, S.Kang, S.T. Lee, Slope stability
Despite FLAC's inherent complexities, it emerges analysis of unsaturated soil slopes based on the
as the more conservative software for slope stability site-specific characteristics: a case study of
studies compared to PLAXIS, because usually gave out hwangryeong mountain, Busan, Korea.
the lowest value for factor of safety in comparison. Sustainability. 12(7), 2839 (2020)
Furthermore, an increase in rainfall intensity results in a https://doi.org/10.3390/su12072839
reduced length of the failure arc (La) and factor of safety 11. K. Terzaghi, Mechanism of Landslides, The
(FOS), with FLAC showing the longest failure length Geological Society of America. Berkey volume,
compared to PLAXIS. This characteristic makes FLAC 83–123 (1950)
a suitable choice for engineers, as it represents the most
12. M.T. Van Genuchten, A closed-form equation for
unfavorable conditions. Further studies and
predicting the hydraulic conductivity of
modification to the DLL file in the PLAXIS program are
unsaturated soils. soil science society of America
needed to simulate the infiltration process.
journal. 44, 892–898 (1980)
https://doi.org/10.2136/sssaj1980.0361599500440
References 0050002x

1. PLAXIS 1D/2D/3D Saturated/Unsaturated Finite

Element Groundwater Modeling (Theory Manual).
(2023)
2. A.S. Chatra, G. R. Dodagoudar, V. B. Maji,
Numerical modelling of rainfall effects on the
stability of soil slopes, International Journal of
Geotechnical Engineering. 13, 425-437 (2019)
https://doi.org/10.1080/19386362.2017.1359912
3. F. Cai, K. Ugai, Numerical analysis of rainfall
effects on slope stability. International Journal of
Geomechanics. 4, 69–78 (2004)
https://doi.org/10.1061/(ASCE)1532-
3641(2004)4:2(69)
4. J.W. Mburu, A-J. Li, H-D. Lin, C-W. Lu,
Investigations of Unsaturated Slopes Subjected to
Rainfall Infiltration Using Numerical
Approaches—A Parametric Study and
Comparative Review. Sustainability. 14, 14465
(2022)
https://doi.org/10.3390/su142114465