Reanalysis of the Kettleman Hills Landfill Slope

Stability Failure
Justyna Adamczyk
AGH-UST University of Science and Technology (Poland)
jadamcz@agh.edu.pl

Abstract The paper presents results of the Kettleman Hills

landfill slope stability failure reanalysis in two dimensional
and three - dimensional models. The analysis was performed
using the computer codes FLAC (v.5.0) and FLAC3D (v.4.0),
respectively, which are Finite Difference Method codes. The
investigation was based on Shear Strength Reduction
Technique (SSR). The approach taken considered the problem
as one of statics. The failure was caused by waste mass
movement along the slip surface located in the geosynthetic
layer, which was characterised by relatively low strength
properties as compared with values of these parameters for the
waste layer as well as for the subgrade. Obtained values for the
factor of safety differ from each other and depend on the kind of
analysis conducted. These results are caused by the possibility of
performing the real geometry of the landfill in 3D modelling
issues.

models. The analysis was performed using the computer

codes FLAC (v.5.0) and FLAC3D (v.4.0), respectively, which
based on the Finite Difference Method. The investigation was
based on the Shear Strength Reduction Technique (SSR). The
approach taken considered the problem as one of statics. The
results of the studies were compared to values obtained in
previous analysis.

I. INTRODUCTION
One of the most important aspects of mining waste
management is its safe disposal. One of the tasks of the MinNovation project has been to promote good practice in this
area. The analysis of landfill failures allows one to better
understand the processes which can occur during their
exploitation. The knowledge gained in this way can be used
to design new solutions, which improve the safety of these
facilities.
The Kettleman Hills landfill is a hazardous waste disposal
facility located on an area of 14.5 ha near Kettleman City
(California, USA). The Kettleman Hills landfill was situated
in flat basin surrounded by side slopes with an average slope
gradient of 23 - 24. The Phase I-A, with an area of about 6
ha and located in northern part of the basin, had been filling
up since 1987. The waste layer was underlined by a protective
compacted liner system. On March 19th, 1988, the maximum
height of the object reached 27 m and the slope stability
failure occurred. The waste layer rapidly slid in a southeast
direction. Landslide dimensions included 10.7 m depth and
4.3 m height. Figure 1 shows the topography of the Kettleman
Hills landfill after the failure occurred. The dashed lines
indicate the location of the landslide. The size of the cracks
created was between 2.5 cm to 120 cm. The largest of these
occurred in the southern part of the landfill. The landslide
also caused disruptions and tears in the liner system on the
landfill side slope.
The Kettleman Hills landfill failure was caused by the
movement of a waste layer, which ran along the slip surface,
located in the protective compacted liner system (double liner
system which had isolated the subgrade) at the interface
between geomembrane compacted clay (on the base liner
system) and geosynthetics layers (on the side liner system).
The paper presents the Kettleman Hills landfill failure
reanalysis in two dimensional and three - dimensional

Fig. 1. The Kettleman Hills landfill topography, after landslide occurred.

Source: Muhsiung Chang, 2005.

II. NUMERICAL MODEL CONSTRUCTION

For the purpose of constructing a two dimensional model
the cross section showed in the figure 2 was used. This
profile is the most representative case for landfill model. The
size of the waste layer includes a height of 27.4 m and a
width of 146.3 m (this profile corresponds to section C
presented in figure 4). This cross-section was directly
implemented in the FLAC v.5.0 code.

Fig. 2. The Kettleman Hills landfill cross-section B-B.

Source: G.M. Filz, J.J.B. Esterhuizen, J.M. Duncan, 2001.

20

Based on the geometry defined in figures 3 and 4, the

three-dimensional model was constructed. Firstly, the
geometry of particular layers was drawn in AutoCAD, then
the surfaces of each layer were created in the Surfer 9.0 code,
afterwards the data were implemented in the FLAC3D code
through a program written in C++. The next step was setting
up a mesh in each model. The generated numerical models
are shown in figures 5 and 6.
The boundary conditions for the bottom and sides of the
models were fixed against horizontal and vertical
displacement, respectively. In both cases, the single interface
between the waste layer and the subgrade was modeled as the
protective compact liner system. The calculations were made
using a Coulomb - Mohr plasticity model.
Fig. 6. The Kettleman Hills landfill 3D model.
Source: Own study.

For the purpose of the analysis three representative layers

were used (waste, interface, subgrade). Table 1 contains
values for geotechnical parameters for each layer which were
implemented for analysis.
TABLE I
GEOTECHNICAL PARAMETERS IMPLEMENTED FOR ANALYSIS
Parameters
Sign
Value
Unit
1730
kg/m3

c
10
kPa
Subgrade

38

E
8x107
Pa

0,25
1570
kg/m3

c
43
kPa
Waste

31

E
2x107
Pa

0,25
17,3
kN/m3

c
10
kPa
Interface

38

K
200
MPa
K0
0,5
Normal stiffness
kn
50x106
Pa/m
Shear stiffness
ks
50x106
Pa/m
Source: Chang M., Mitchell J.K, Seed R.B. Model Studies of the 1988
Kettleman Hills Landfill Slope Failure, Geotechnical Testing Journal, 1999,
pp. 61 66
Layer

Fig. 3. The Kettleman Hills landfill basin geometry (top view).

Source: Muhsiung Chang, 2005.

III. LANDFILL SLOPE STABILITY ANALYSIS (SSR)

The most likely explanation is that the damage occurred in
the liner system, which underlined the waste layer. This
assumption is confirmed by observations, which were made
after the failure had happened as well as the results of the
previous stability analysis.
The Kettleman Hills landfill slope stability failure analysis
was conducted using the reduction of the friction angle value
for the interface.
In the two-dimensional slope stability analysis of the
Kettleman Hills landfill, the instability occurred for the
factor of safety FS = 0.99 (for the friction angle = 4.3).
Figure 7 illustrates the distribution of the shear strain in the
layer of waste and its relationship to the factor of safety
value. It can be seen that shear strains are growing to a
decreasing value of FS. Damage of the waste layer appeared

Fig. 4. The Kettleman Hills landfill basin geometry (cross-sections).

Source: Muhsiung Chang, 2005.

Fig. 5. The Kettleman Hills landfill 2D model.

Source: Own study.

21

first in its lower part in close proximity of the interface.

Afterwards the failure embraced the middle part of the layer.
At the moment of instability the damage was present
throughout the layer.
The value of the friction angle received for instability in
this analysis is comparable with the results obtained in other
studies for the facility. M. Chang in his research for a 2D
model analysis gained a minimum value of = 5.0 .
The value for the friction angle for which there was the
instability of the slope landfill in 2D analysis ( = 4.3) was
the starting point for the continuation of a three-dimensional
model calculation.

Table 2 shows the successive changes in the value for the

friction angle, which were applied to the 3D model in order to
determine the strength parameters for interface at the time of
failure.
TABLE II

FRICTION ANGLE VALUES FOR INTERFACE (3D ANALYSIS)

4,3
6,0
8,0
9,0
[]
FS
0,76
0,9
0,99
1,05
Source: Own study

Further calculations carried out for the 3D model showed

that the loss of stability was obtained for a friction angle
of >> 8.0. Figure 8 shows the increase of the deformation
connected with decreasing of the friction angle value. The
figure includes also the comparison between shear strain and
displacement distribution which arise for the given value of
the friction angle. It can be seen that displacements
propagation is compatible with shear strain distribution. It is
visible particularly, when the factor of safety value dropped
under unity.
a

Fig. 6. Factor of Safety and Shear Strain distribution in 2D analysis.

Source: Own study.

Three dimensional analysis showed that for the value of the

friction angle = 4.3 the factor of safety is much lower than
in 2D results, and it is set to FS = 0.76 (Figure 7). This
indicates that landfill slope stability was lost at a higher value
of the friction angle than the one received in 2D analysis.
The stress strain distribution obtained from analysis which
are shown in figure 7 corresponds to disruptions and cracks
observed on the landfill after the failure occurred.

Fig. 8. Shear strain distribution (a); Displacement distribution (b).

Source: Own study.

Fig. 7. Shear stress increase - 3D analysis.

Source: Own study.

By comparing 2D and 3D analyses, it was observed in both

cases that shear strain appeared first at the bottom of the
waste layer (in the close neighborhood of the interface). In

22

other words, the slippage in the 3D model occurred at first

along the back part of waste layer (FS values from 1.05 to
0.99) - just like in 2D analysis results. Afterwards, when
shear strain zone is increasing to the top of the waste layer,
the value of the factor of safety is falling to FS = 0.99. At the
point of instability occurred the shear strain zone, which
included all cross-sectional thickness of waste layer. Using
low-strength parameters for the interface, for which the factor
of safety is lower than unity, one is able to determine the
progress of the slip surface. For the value of friction angle
= 6.0 the failure continued to develop by crossing again
the entire waste layer up to the upper corner, which connects
the waste layer with interface (FS=0.9). Continued damage
progress (for = 4.3) caused the sliding surface at the back
side of interface. The shape obtained for FS = 0.76 recalls a
characteristic wedge figure.
Further comparison between 2D and 3D analysis results
indicates that for the same value of friction angle = 9.0 we
received a different value for the factor of safety - FS = 1.31
for 2D, FS = 1.05 for 3D (figure 9). The lower value of the
FS for 3D analysis is caused by the possibility of reproducing
the real geometry conditions of the object.
The next indicator for the coming landslide is a large
increase of displacement velocities. Table 3 compares
displacement velocity for different factor of safety values. It
can be seen that when the FS value is falls, the movement of
the velocity vector becomes much more orderly.

The abnormal displacement increase is related to the values

of the velocity and it is also the indicator for impending
landslide. Figure 10 shows increasing displacement obtained
in 2D analysis which are connected with decreasing values of
factor of safety.
As in the case of the velocity displacement vector, there is
a direct connection between the falling value of factor of
safety and spreading displacements.

Fig. 10. Abnormal displacements 2D analysis.

Source: Own study.

Landslide development can be seen also in plasticity index

distribution. Figure 11 presents the failure zone spreading as
the result of shearing, which is due to the reduction of the
mechanical property values. This destruction process starts in
the lower part of the waste layer, afterwards cuts the layer and
finally reaches the entire cross section of the layer and passes
to the damage zone due to tensile stress.
The progressive failure effects were more significant in
interface plasticity analysis. Figure 11 shows the relation
between slip surface development and decreasing factor of
safety values. For the factor of safety is FS = 1.05, the slope
is stable because the interface is not yet fully plasticized. The
figure shows the place in the basin of the landfill, on the
interface surface, which has not yet reached the peak shear
strength. The same location was received by Filz et.al. (2001)
in the analysis of landfill slope stability (the analysis was
carried out using the finite element method).

TABLE III
VELOCITY DISPLACEMENTS 2D ANALYSIS

FS
1,31
0,99

Velocity (m/s)
8,376e-5
2,354e-4

Source: Own study

Fig. 9. Comparison between 2D and 3D analysis results.

Source: Own study.

Fig. 11. Plasticity index distribution - 2D analysis.

Source: Own study.

23

IV. CONCLUSIONS

Figure 12 shows a progressive slip surface plastification

and its relation to a decreasing factor of safety value. For the
friction angle = 9.0, the slip surface area is not yet fully
plasticized. It can be seen that in the middle of the basin there
is a place, which has not reached the peak shear strength
value. The location of this place is in agreement with that
received by Filz et. al. (2001).

The stability of the Kettleman Hills slope failure was

reanalyzed using 2D and 3D models implemented
in FLAC v. 5.0 and FLAC 3D v. 4.0, respectively. The
calculations relied on the Finite Difference Method The
analysis was conducted using the Shear Strength Reduction
Technique. In the present study, the parameter value which
decreased was that of the friction angle.
The 3D model successfully reflected the landslide
geometry. The trend of damage in the 3D model was located
in the same places as real cracks which were observed after
failure occurred.
The slip surface was located in a geosynthetic layer which
was characterized by relatively poor strength values in
comparison with the waste layer and subgrade parameters.
It was found that the failure occurred when the value of the
friction angle amounts to 4.3 (for 2D analysis) and 8.0 (for
3D analysis). Similar results for the stability analysis were
received by Chang.
The analysis also made it possible to obtain information
about the progress of the landslide, including its direction,
range and slip surface shape.
The considerations above lead to the following
conclusions:
1. The key point for the Kettleman Hills landfill slope
stability failure was insufficient strength of the
geosynthetic layer (complete interface plasticity was
equivalent to the slide surface occurrence).
2. Comparison of the results of the planar and the spatial
analysis shows that there are notable differences
between the values of the friction angle for which the
landfill slope instability occurred.
3. The reason that the spatial analysis registers instability
for greater values of the friction angle is connected
with the ability to reproduce real conditions of the
landfill geometry as well as to perform propagation of
the shear strain deformations.
4. Three-dimensional stability analysis confirms the
major benefits of using spatial analysis. The
continuation of this type of calculation allows one to
determine the appropriate size of the strength
parameters for materials used for protective and
isolate barriers in compacted liner systems used in
landfill construction.
ACKNOWLEDGMENT
Financial support for this work was given by organisations
named below:

Part-financed by the European Union (European Regional Development

Fund and European Neighbourhood and Partnership Instrument)
Fig. 12. Plasticity distibution on the interface - 3D analysis.
Source: Own study.

Scientific work financed from science funds in 2011-2013 allocated for cofinancing implement international project

24

REFERENCES
1.

2.

Caa M., Flisiak J., Tajdu A. (2006), Slope stability

analysis with FLAC in 2D and 3D, FLAC and
Numerical Modeling in Geomechanics - 2006,
edited by P. Varona, R. Hart, Itasca Consulting
Group, Inc., Minneapolis, pp. 11-14.
Chang M. Three dimensional stability analysis of
the Kettleman Hills landfill slope failure based on

3.

4.

25

observed sliding block mechanism, Computers

and Geotechnics 32 (2005) pp. 587 599
Chang M., Mitchell J.K, Seed R.B. Model Studies
of the 1988 Kettleman Hills Landfill Slope Failure,
Geotechnical Testing Journal, 1999, pp. 61 66
Filz G., Esterhuizen J.B., Duncan J.M. Progressive
Failure of Lined Waste Impoundments, Journal of
Geotechnical and Geoenvironmental Engineering,
October 2001, pp. 841 848