See discussions, stats, and author profiles for this publication at: [Link]

net/publication/257564345

Stress–Strain Analysis of Buckling Failure in Phyllite Slopes

Article  in  Geotechnical and Geological Engineering · February 2012

DOI: 10.1007/s10706-012-9556-8

CITATIONS READS

6 137

2 authors:

Luana Cláudia Pereira Milene Sabino Lana

Universidade Federal de Viçosa (UFV) Universidade Federal de Ouro Preto
6 PUBLICATIONS   8 CITATIONS    36 PUBLICATIONS   57 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Modelling of granular flow View project

Study of rupture mechanisms in small rock slopes View project

All content following this page was uploaded by Milene Sabino Lana on 28 January 2019.

The user has requested enhancement of the downloaded file.

Stress–Strain Analysis of Buckling Failure
in Phyllite Slopes

L. C. Pereira & M. S. Lana

Geotechnical and Geological

Engineering
An International Journal

ISSN 0960-3182

Geotech Geol Eng

DOI 10.1007/s10706-012-9556-8

1 23
Your article is protected by copyright and all
rights are held exclusively by Springer Science
+Business Media Dordrecht. This e-offprint
is for personal use only and shall not be self-
archived in electronic repositories. If you
wish to self-archive your work, please use the
accepted author’s version for posting to your
own website or your institution’s repository.
You may further deposit the accepted author’s
version on a funder’s repository at a funder’s
request, provided it is not made publicly
available until 12 months after publication.

1 23
 Author's personal copy
Geotech Geol Eng
DOI 10.1007/s10706-012-9556-8

TECHNICAL NOTE

Stress–Strain Analysis of Buckling Failure in Phyllite Slopes

L. C. Pereira • M. S. Lana

Received: 10 October 2011 / Accepted: 14 September 2012

 Springer Science+Business Media Dordrecht 2012

Abstract Buckling failure occurs in the open pit Keywords Slope stability  Buckling failure 
mines in the Quadrilátero Ferrı́fero of the State of Minas Numerical models  Stress–strain analysis
Gerais, Brazil. In this study, this failure mechanism is
evaluated through stress–strain analysis, using the finite
element software named Phase2. Many representative
hypothetical models of the slopes were created in order
to reproduce the conditions in open pit mines, like their 1 Introduction
geometry and foliated joints with kinematic conditions
for buckling failure. Elastic and plastic simulations have There is one important iron and gold province, called
been done. Strength and deformability parameters, such Quadrilátero Ferrı́fero (QF) in the State of Minas Gerais,
as discontinuity stiffness and cohesion, as well as the Brazil. Its geology is very complex and many kinds of
in situ stress field, are analyzed in order to understand different slope failures, involving structural discontinu-
their influence in the buckling failure mechanism in the ities, are common in this region. According to Diláscio
global open pit mine slopes excavated in phyllites. As (2004), a major part of the ore footwall in mine slopes at
studies about this problem are extremely rare, the QF is compounded with phyllites of the Itabira and
present research brings important contributions to Caraça Geologic Groups, in association with schists of
establish the basic conditions for this kind of failure. the Nova Lima Group. As the Iron Formation is a large
aquifer, the rocks have been weathered, even at great
depths, creating low-strength rock masses. Rock slopes
in this material can present different kinds of failure. The
waste slopes in the open pit gold mines are compounded
L. C. Pereira  M. S. Lana
Graduate Program in Geotechnical Engineering—NUGEO, with schists of the Nova Lima Geologic Group. In this
Federal University of Ouro Preto—UFOP, case, the mine slopes can present stability problems,
Campi Morro do Cruzeiro, Bauxita, Ouro Preto, such as flexural toppling and buckling through foliation
Minas Gerais CEP: 35400000, Brazil
discontinuities.
e-mail: luanac_pereira@[Link]
In the QF some mine waste overall slopes have
M. S. Lana (&) presented buckling failure, however this mechanism is
Mining Engineering Department—DEMIN, Federal not completely understood. Occurrences of buckling
University of Ouro Preto—UFOP, Campi Morro do
have been observed in slopes of two mines, at Córrego
Cruzeiro, Bauxita, Ouro Preto, Minas Gerais CEP:
35400000, Brazil do Sı́tio Mine and at Pau Branco Mine, both in phyllite
e-mail: milene@[Link] rock mass.

123
 Author's personal copy
Geotech Geol Eng

In Córrego do Sı́tio gold mine buckling occurred as tension cracks in an extent of 400–500 m beyond the
a local failure mechanism in a bench slope (Fig. 1). slope crest during buckling development.
Lopes (2006) demonstrated that this failure mode has Folding of the foliation, with the consequent
had no influence in global slope stability. This author verticalization of this structure has created the favor-
studied a major failure that occurred in the same slope able kinematic conditions for buckling occurrence.
through numerical analysis; the results showed that the Changes due to folds are responsible for the winding
failure mechanism was a circular surface combined pattern of the layers and for relatively abrupt changes
with plane shear failure. in the foliation dip angle. This pattern can be seen in
A case of particular interest is the buckling failure Fig. 4.
which occurred at Pau Branco iron mine, as it reached In order to improve the understanding of buckling
many benches of a phyllite rock slope. As it can be failure, some hypothetic slope models based on typical
observed in Fig. 2, a huge portion of the slope is QF phyllites were studied in this work, using stress–
subjected to large displacements due to buckling. This strain analysis. The finite element software called
failure is studied by Silva (2010) through numerical Phase2, of Rocscience, Canada, was used. According
models, to permit the comprehension of the failure to Sjöberg (1999), stress–strain analysis permits the
process. Also the carrying load conditions involved study of complex mechanism failure with significant
and the strength and deformability parameters of the deformability that involves intact rock failure or a
rock mass and discontinuities were determined by combination of failure mechanisms involving discon-
back-analysis. tinuities and intact rock. Besides this, it is possible to
Buckling failure at Pau Branco occurred in 2002, identify the factors that cause the failure.
affecting many benches. A geological preliminary Parametric analyses have been made with a wide
model to understand the phenomenon, proposed by a variety of input data, like the slope height, the in situ
consulting company (BVP Engenharia 2007), was stress state, the rock strength and others. The goal was
used in the analyses. This model is represented by a to study the influence of the input data in buckling
typical geological section (see Fig. 3). The foliation failure development in overall mine phyllite slopes.
trace of the rock mass in Fig. 3 shows the dip The strength and the deformability input data of rock
variability of this structure, due to folding. BVP material have been defined considering previous
Engenharia (2007) reported a region of large displace- studies in QF. These studies include back-analysis
ments due to buckling (Fig. 3); they reached approx- and shear tests, presented in Lopes (2006) and Lopes
imately 3.5 m. BVP also reported the occurrence of et al. (2007).

Fig. 1 Buckling failure in a

slope bench of a gold mine
in QF

123
 Author's personal copy
Geotech Geol Eng

Fig. 2 Buckling failure observed in an iron open pit mine, in QF

Fig. 3 Proposed model for

buckling failure in Pau
Branco (BVP 2007)

The major aim of this research is to establish typical geomechanical properties. The major kinds of failures
conditions for buckling failure in large scale slopes are: planar, wedge, circular, toppling and buckling.
containing phyllites. Some of the input data is difficult According to Giani (1992), buckling failure can
to obtain, such as discontinuity stiffness; they must be happen when there is a thin slice in the rock mass of
established by back-analysis during the numerical the slope, due to the presence of discontinuities nearly
analysis. parallel to the slope’s face. Another important condi-
tion refers to discontinuity dip; it needs to be equal to
or higher than the slope dip.
2 Buckling Failure Buckling is common in stratified sedimentary rocks
where bedding planes separate slabs; it can be induced
A rock or land slope can present different kinds of by external forces, like, water pressure, plate weight, or
failures depending on its geological, geometrical and stress concentrations on the plane of the plates

123
 Author's personal copy
Geotech Geol Eng

Fig. 4 Phyllite slabs folded showing favorable conditions for buckling

(Diláscio 2004). This type of failure can also happen in face angle, / is the discontinuity friction angle, wp is
metamorphic rocks, for example, in phyllites where the discontinuity dip.
cleavage or schistosity is significant and regular (Froldi According to Sjöberg (1999), buckling failure is
and Lunardi 1995); as it has been presented before. due to high axial stress in the rock slab, when its
Buckling failure happens when there are favorable thickness is very small regarding to its length.
kinematic conditions (Eqs. 1, 2). Gradual tension Kutter referenced by Froldi and Lunardi (1995)
crack formation and buckling only in the lower part considered that many factors can affect buckling:
of the slope is very common (Fig. 5).
• the presence of subparallel joints cutting the lower
The kinematic conditions for buckling with zero
dip slab;
discontinuity cohesion and in absence of water pressure
• the presence of waviness and roughness in main
are:
discontinuities;
ap  h  / • the shear strength and previous sliding along
ð1Þ
wp  h discontinuities.

where ap is the discontinuity pole plunge, h is the slope Cavers (1981) established the conditions for buck-
ling failure, based on the Euler Theory. He presented
Euler’s expression for the maximum load per unit
slope width before buckling occurs. In order to apply
Euler’s concepts to slopes, he assumed that only a
certain portion of rock layer near the slope toe region
buckles; the rest of the rock layer only provides an
axial load to this buckling portion. Base friction model
tests results gave a range for the ratio of the buckling
slab length to the total slab length equal to 0.36–0.46.
The lower bound of this ratio (0.36) is for a slab
slightly too thick to buckle and the upper bound (0.46)
is for a slab thinner than the one required to buckle.
Fig. 5 Buckling model suggested by Sjöberg (1999) Cavers assumed that the ratio of the buckling slab

123
 Author's personal copy
Geotech Geol Eng

length to the total slab length is 0.5, which is slightly

conservative compared to the results of base friction
models.
Adhikary et al. (2001) rearranged Caver’s expres-
sion to yield the following expression for a non-
dimensional term named the critical loading parameter
(cH3/Eh2) for flexural buckling:

cH 3 p2
1
2
¼ sin a  cos a tan /j ð2Þ
Eh 2:25

where c is the unit weight of the rock layer, E is the

effective elastic modulus in the direction of layering,
H is the slope height, h is the layer thickness (joint
spacing), /j is the joint friction angle, a is the slope Fig. 6 Designed chart for foliated rocks slopes against flexural
angle. buckling failure for the case of joint friction angle of 0, 10, 20,
Adhikary et al. (2001) commented on that the fixed 30 from Adhikary et al. (2001)
length of rock layer subjected to buckling proposed by
Cavers imposes a very strict constraint in his model necessary to analyze the possibility of other types of
formulation. These authors considered that it is more failures. Simple calculations can be done to examine if
natural to assume that this length depends on the slope shear failure in rock mass may precede buckling.
angle. Therefore, Adhikary et al. proposed a compre- According to Adhikary et al. a conservative estimate
hensive model of flexural buckling where the buckling of the slope height for shear failure can be calculated
layer length proposed by Cavers is not an input by Eq. 3. This estimate corresponds to failure under
parameter but it remains a function of the slope axial loading, when the slab weight is higher than the
geometry. This approach is a large deformation model uniaxial compressive strength.
on the basis of Cosserat theory (1909). Their results
were implemented in finite element method of AFE-
NA software.
On the basis of their numerical calculations,
Adhikary et al. proposed a chart for the design of
slopes in foliated rock masses subjected to buckling
(Fig. 6). This chart relates the slope angle (varying
from 60 to 90), the critical loading factor proposed in
the model (varied of 0.1–100) and four joint friction
angles (0, 10, 20 and 30). However this chart does not
consider the possibility of the existence of water
pressure in the joint or cross-joints (Adhikary et al.
2001).
Figure 7 relates slope angle to the critical layer
thickness or critical joint spacing, calculated for
typical values of E = 50 GPa, c = 25 kN/m3, slope
heights of 100, 200 and 400 m, and joint friction angle
equal to zero. For example, buckling failure can
happen at a slope angle of 80 and height of 200 m or
more, when joint spacing is \0.4 m (Adhikary et al.
2001).
Fig. 7 Critical layer thickness versus slope angle for slope
The critical loading factors can be higher for gentle heights of 100, 200 and 400 m (E = 20GPa, c = 25 kN/m3,
slopes in comparison to vertical ones. So it is / = 0; Adhikary et al. 2001)

123
 Author's personal copy
Geotech Geol Eng

rc impossible to introduce foliated discontinuities in the

Hshear ¼ ð3Þ
c entire model because the computation effort is so large
that program execution is not viable. So, 21 disconti-
where Hshear is the critical slope height for shear
nuities were included in the model behind and at the toe
failure, rc is the uniaxial compressive strength of rock
of the slope face. This model configuration was reached,
material, c is the unit weight of the intact rock layer
after many tries, as representative of buckling failure in
material.
QF and the computation effort is fair; some models
The height Hshear must be greater than Hb (critical
required a long execution time, even the elastic ones.
height for buckling) for the possibility of any type of
The slope heights were considered as input param-
buckling failure to occur. The critical height Hb can be
eters. They were established based on the typical slope
calculated as follows:
heights encountered in large open pit mines. The joint
 1=3 dip is 63; this value was established by kinematic
AEh2
Hb ¼ ð4Þ analysis (Eq. 1). The slope face angle is 60 and the
c
friction angle is 30, so the discontinuity dip must be
where A is the critical loading factor that can be equal to or [60. The established value was 63.
obtained from chart of Fig. 6, for a particular slope The chart in Fig. 6 was used to determine the
angle. critical loading parameter. As a result, the layer
The combination of Eqs. 3 and 4 leads to an thickness (joint spacing in the model) was calculated:
inequality that is necessary for any buckling failure • for slope height of 100 m, the joint spacing was
before shear failure occurs, Eq. 5. 0.3 m;
 1=3 • for slope height of 200 m, the joint spacing was
AEh2 rc 0.75 m;
Hb ¼ \ ð5Þ
c c • for slope height of 300 m, the joint spacing was
Equation 5 can be arranged to obtain an expression 1.4 m;
of a limiting layer thickness, below which buckling • for slope height of 400 m, the joint spacing was
failure for a particular slope occurs, if the critical 2.1 m.
height is attained (Eq. 6). • for slope height of 500 m, the joint spacing was
  3 m.
1 rc 2 rc 
h2 \ ð6Þ Displacement boundary conditions were used. Null
A c E
horizontal displacements were used in lateral bound-
3 Numerical Analysis aries and null vertical displacements in lower bound-
ary (Fig. 8).
A numerical model with plane strain conditions was The model dimensions were: 6 times the slope
analyzed in this study, using Phase2 from Rocscience height for the vertical and nearly 9 times the slope
Inc. (2008) (Canada), in order to understand the height for the horizontal dimension (Fig. 8). Sjöberg
conditions involved in the buckling failure of foliated (1999) suggests model dimensions equal 2–3 times the
rocks for large slopes. Some parametric analyses have total slope height for the vertical dimension and 3–4
been done changing the field stresses and geomechan- times the excavated width for the horizontal dimen-
ical properties of the rock material. sion. This suggestion was not used, because it led to
Many hypothetical various possible models of mine the occurrence of unacceptable displacements at the
overall open pit slopes were constructed. The slope boundaries of the model. The dimensions in the model
heights were 100, 200, 300, 400 and 500 m and the were established by trial and error, until there had been
slope angle was 60. no significant displacements at the model boundaries,
Main foliation discontinuities were explicitly intro- as displacements at the boundaries were fixed accord-
duced in the model as joint elements, as they cause the ing to boundary conditions applied in the model. The
buckling failure. The foliated discontinuities were put in displacements were restrained along x direction in the
a particular area of interest, near the slope face and along lateral boundaries and along y direction in the bottom
the whole vertical extension of the model. It is boundaries of the model.

123
 Author's personal copy
Geotech Geol Eng

Fig. 8 Model with finite element mesh

The model was built in two stages. In the first stage, lithostatic rock column weight and the horizontal
the model represents the original situation, before in situ stress varied between 0.25 and 3.0 times the
slope excavation; and in the second stage, the slope vertical in situ stress, a range with very wide values. It
was excavated. The slope was excavated in only one is quite difficult to have in situ horizontal stress
stage as the purpose of the excavation is only to get the beyond this range.
correct values of the stress field. Rock material was considered to be linear elastic, in
The finite element mesh was established by trial order to obtain the observed typical buckling behavior
and error, using improved mesh resources offered in in QF, such as the presence of tension cracks (zones
Phase2 (Fig. 8). There are more discretizations in the with negative values of r3) in the slope crest, without
region of interest, close to the discontinuities. much computational effort.
In order to simplify the model, the entire rock mass After elastic simulations, plastic models were used;
consisted of phyllite. Although this is a crude simpli- changing the rock mass and joint cohesion to study the
fication it is valid as the region of interest is restricted to effect of these parameters on the failure mechanism.
the slope face proximity. Besides this it avoided the Nearly 170 models were built and analyzed.
estimation of many different material properties, which
could influence the results. The input properties (Pois-
son ratio, cohesion, friction angle, tension strength, 4 Results
deformability parameters and others) of phyllite and the
joints were established based on technical literature The results and their analyses were divided into two
about QF, such as in Lopes (2006). During the study, groups, the elastic and plastic models.
parametric analyses were made, adjusting the initial
properties into a range of adequate values for QF, based 4.1 Elastic Models
on reports of consultants or scientific studies.
Parametric analyses for field stress were also Although the intact rock does not fail in the elastic
performed. The vertical in situ stress was related to model, since stress redistribution due to yielding is not

123
 Author's personal copy
Geotech Geol Eng

possible, this model can provide important informa-

tion about joint failure mechanisms with a smaller
amount of input data than the plastic model.
The following input data was used for the phyllite:
Poisson’s ratio 0.2; Young’s modulus 3,030 MPa;
tensile strength 0.055 MPa; friction angle 33; cohesion
0.15 MPa; unit weight 0.02 MN/m3. For discontinu-
ities: tensile strength equal to zero; cohesion 0.16 MPa;
friction angle 30; normal and shear stiffness of 50 and
5 MPa/m, respectively. These data were considered as
typical of phyllites in QF; they were taken from
technical reports and scientific studies, such as Lopes
(2006).

4.2 Influence of the In Situ Stress Field

Fig. 9 Variation of k according to depth (Brady and Brown
The vertical in situ stress was considered equal to the 2004)
unit weight of the rock multiplied by the depth below
ground surface. Horizontal stresses were calculated
using the ratio k of horizontal–vertical stress. • the total displacement; and
The first assumption for k in the models was that • the contour values of strength factors (SF).
there was complete lateral restraint, a condition for
A value of k = 0.25 was excluded from the analysis.
gravitational loading.
This value did not lead to the occurrence of a tension
g crack region in the slope crest for the models of 100 and
k¼ ð7Þ
1g 200 m high (Fig. 10); this fact was observed in the
typical buckling occurrences in QF, as in Pau Branco
where k is the ratio of in situ horizontal to vertical
Mine. Also the slopes with 300, 400 and 500 m high
stress, g is the Poisson’s ratio.
with a k value = 0.5 did not lead to the occurrence of a
As the Poisson’s ratio was 0.2, k was equal to 0.25.
tension zone in the slope crest (Fig. 11), although in
The other assumptions of k were: 0.5; 1.0; 1.5; 2.0;
smaller slopes such zone was observed, Fig. 12. In
2.5 and 3.0. These values were based on the chart
Figs. 10, 11 and 12, tension cracks are represented by
suggested by Brady and Brown (2004), where k varies
zones with negative values of r3. In Fig. 12 the length
according to a function of the depth (Fig. 9). This chart
of the region of r3 negative values beyond the slope
was established by nearly 900 in situ stress measure-
crest is comparable to the field evidences of tension
ments made for many different mining, civil engi-
cracks extent related to buckling in Pau Branco Mine.
neering and petroleum projects.
For slopes of 100 m height, the tension region has
The following specific factors were chosen to
maintained almost constant for different values of k. For
interpret the model results originated by the in situ
other slopes, this region increased with the value of k.
stress variation. They are:
All the models, with different values of height and
• the standard of buckling in phyllites in QF, like the k, presented buckling deformability near the slope toe
presence of tension cracks in the slope crest and the (Fig. 13). This result was expected, because the slopes
typical buckling deformability of the layers. This meet the kinematic conditions for buckling.
fact must be reproduced in the model, as it has been Maximum total displacement occurred in the slope
observed in the field. The presence of tension face or in the slope base. When maximum displacement
cracks was associated to the occurrence of tension values occurred in the slope face, where the disconti-
(negative values of r3); nuities have been inserted, buckling deformability is
• the vector displacement field, magnitude and very well-defined, as showed in Fig. 13a. High values
direction; of displacements can be seen close to the slope toe, in

123
 Author's personal copy
Geotech Geol Eng

Fig. 10 Model with a slope height of 100 m and k = 0.25

Fig. 11 Model with a slope height of 500 m and k = 0.5

buckling region. Buckling occurs near the slope toe and When maximum total displacement occurred in the
the displacements decrease towards the slope base. slope base, typical buckling deformability is affected by
Maximum displacement in Fig. 13a is equal to 5.3 m. these high values of displacements in the slope base, as
This behavior is observed for all models when showed in Fig. 13b. Buckling deformability pattern is less
k = 1.5–3.0. prominent than in the previous case, as displacements

123
 Author's personal copy
Geotech Geol Eng

Fig. 12 Model with a slope height of 100 m and k = 0.5

increase towards the slope base and reach its maximum The values of k = 2.0, 2.5 and 3.0 presented large
value in front of the discontinuity region. Maximum regions of low SF, indicating a big yield-potential
displacement in Fig. 13b is equal to 4.2 m. This behavior region. Typical buckling occurrence in open pit mines
is observed for all models when k is below 1. of QF did not lead to very large failures, so this fact
There is a transition point for k = 1 with well- permits the interpretation that these values of k are not
defined buckling deformability but maximum total typical of the QF, but this affirmation needs to be
displacement in the slope base, as showed in Fig. 14. confirmed by results from plastic analysis.
Maximum displacement in Fig. 14 is equal to 8.3 m.
In terms of maximum total displacements, it was
observed that the higher the slope, the greater the 5 Joint Stiffness Effects
displacement was for the same k; and the larger the
k value, the greater the displacement was for the same Stiffness is a joint deformability parameter that relates
height (Fig. 15). the stress and the displacements in normal and
The software Phase2 allows the calculation of the tangential directions. It is expressed in terms of normal
SF values, which is the ratio between the critical stiffness (kn) and shear stiffness (ks). It is difficult to
stresses reached according to a strength criterion by obtain reliable estimates for these parameters and the
the mobilized stresses. The SF was used as a first possible range of stiffness values is very broad.
indicator of yielded regions, to be confirmed in plastic Bandis et al. (1983) suggested that the ratio kn/ks
models. The region with low SF (between 0 and 1) varies with the normal stress along the joint (rn ):
increased by increasing the value of k; this behavior
• for rn B 0.01 MPa ) kn = 100ks;
was observed for all heights (Fig. 16).
• for ) C 0.01 MPa ) kn = 10ks.
Finite element solutions for models of slope height
of 500 m and k = 2.5 and 3.0 did not converge. Such These relations were used for the models.
situation was interpreted as a global collapse of the It has been expected that, for high values of
slope. Failure is related to buckling phenomena stiffness, buckling would not occur, but this fact was
because of large displacements due to discontinuities, not confirmed by the models. Buckling failure behavior
as an elastic behavior was assumed for the rock mass. was the same for all the models; the only distinct effect

123
 Author's personal copy
Geotech Geol Eng

(a)

(b)
Fig. 13 Models with a slope height of 200 m showing deformation contours and deformed mesh. a k = 1.5, b k = 0.5

was the decrease in displacements. For the largest were analyzed. The goal was to understand how the
stiffness values (50/500 MPa/m), there was a maxi- discontinuities affect the model.
mum displacement of 0.7 m (Fig. 17). Buckling failure causes a tension region in the slope
crest and buckling deformability near the slope toe. A
continuous rock mass does not exhibit this behavior, as
6 Discontinuities Effects confirmed by the model results. Besides, great
displacements happened because of discontinuities,
Some models without any discontinuity, for slope as shown in Fig. 18, with the maximum displacement
heights of 100 and 200 m and different values of k, being 1.80 m, in the slope face. In the continuous

123
 Author's personal copy
Geotech Geol Eng

Fig. 14 Model with a slope height of 300 m, k = 1, showing deformation contours and deformed mesh

model, this displacement reached 0.22 m near the

slope toe (Fig. 17).
Displacement (m)

7 Plane failure

Models where plane failure is kinematically possible

were analyzed for slope heights of 100 and 200 m.
Plane failure can occur when the discontinuities are
nearly parallel to the slope face. The discontinuities k
must have a smaller dip angle than the slope face
angle; the discontinuity dip must be higher than its
friction angle, for a discontinuity with zero cohesion, Slope Height (m)
and release lateral surfaces must be present. The
Fig. 15 Increase of the maximum total displacement with
kinematic conditions in absence of water pressure and increase of height and k
for zero discontinuity cohesion are:
wp  /
ð8Þ
wp \h 7.1 Plastic Model

where h is the slope face angle, / is the discontinuity In plastic models, failure can be simulated because
friction angle, wp is the discontinuity dip angle. when the limit stresses are reached, a process of stress
Slope face angle of the model was 608 and friction redistribution takes place; plastic points can become
angle 30, so the discontinuity dip must be [30 and elastic or yield.
\60. The assumed discontinuity dip value was 50. Plastic analyses are more representative of rock
Results of simulation showed that the tension mass behavior. However, it requires knowledge of the
region beyond the slope crest was smaller in the plane peak and residual strength parameters of the material,
failure models than in the buckling models (Figs. 19, and especially in case of the residual ones it is difficult
20). Besides, displacement was larger in plane failure. to obtain reliable estimates.

123
 Author's personal copy
Geotech Geol Eng

Fig. 16 Low SF region for a slope height of 300 m and k = 3.0

Fig. 17 Slope height of 100 m, k = 1.5, without discontinuities

The elastic model input data is the same for plastic residual friction angle is 25 and the residual cohesion
models, except for the residual rock mass parameters. is 0.03 MPa.
These were based on internal reports of Pau Branco In plastic models, some parameters were chosen to
Mine (Vallourec and Mannesman Company). The interpret the results, such as:

123
 Author's personal copy
Geotech Geol Eng

Fig. 18 Slope height of 100 m, k = 1.5, with discontinuities

Fig. 19 Section with plane failure, slope height of 200 m, k = 2.0 and tension region

• buckling failure pattern; Plastic simulations showed that values of k = 2.0,

• vector displacement pattern; 2.5 and 3.0 are not applicable to QF, because the
• total displacement; model has a very large plastic area. In some analysis,
• tension and shear plastic zones. the plastic zones extended to the entire model, when

123
 Author's personal copy
Geotech Geol Eng

Fig. 20 Section with buckling failure, slope height of 200 m, k = 2.0 and a greater tension region than plane failure

Fig. 21 Slope height of 100 m and rock material cohesion equal to 0.15 MPa, k = 3

k is = 3.0 (Fig. 21). This result has been predictable others, permit to infer that adequate values of k which
by elastic simulations through regions with a low SF. represent QF conditions for buckling failure occur-
Results of elastic and plastic analyses, such as rence are 1.0 and 1.5. Lopes (2006) has also found a
plastic region, buckling pattern, displacements and value of k = 1.5 in back analysis of Corrego do Sitio

123
 Author's personal copy
Geotech Geol Eng

Fig. 22 Slope height of 100 m, k = 1.5, rock material and foliation cohesion equal to 0.2 MPa

Fig. 23 Slope height of 100 m, k = 0.5, and cohesion of 0.3 MPa

Mine case study. Silva (2010) has found that a value of The model for a slope height of 200 m and k = 1 is
k = 1 is typical in back analysis of Pau Branco Mine comparable to the model studied by Silva (2010), for
case study. Anyway, this suggestion should be buckling in Pau Branco Mine, which had the same
confirmed by field observations. slope height and k value. Both models went into

123
 Author's personal copy
Geotech Geol Eng

Fig. 24 Slope height of 100 m, k = 1.5 and cohesion of 0.3 MPa

collapse and presented a combined failure mode, All models of 300 m high slopes went into collapse,
buckling along discontinuities and circular failure in indicating a global failure of the rock mass. Combined
the rock material. failure (buckling and intact rock failure) has been
observed for k values of 0.5, 1.0 and 1.5, as well as for
all cohesion values. Intact rock failure has not
decreased with the increase of k, as in case of 100 m
8 Cohesion Influence high slopes. So, it can be concluded that mine
deepening will cause a great impact on slope stability.
Plastic models with different cohesion values for rock
material and foliation of 0.05, 0.15, 0.2, 0.3 and
0.4 MPa were analyzed. Models with slope heights of 9 Conclusions
100 and 300 m were used.
All the models with a slope height of 100 m, and The geotechnical study permitted a good comprehen-
k = 1.5 showed maximum displacement values of sion of the buckling failure mechanism in phyllite
6.0 m (Fig. 22). slopes of open pit mines at QF. The results permit the
There has been combined failure (buckling and prediction of this mechanism for similar situations in
intact rock failure) for k values = 0.5 and 1.0; and for the QF.
all cohesion values and slope height of 100 m. Intact The results are very similar to buckling occurrence
rock failure decreased with increasing k values in Pau Branco Mine. Buckling pattern deformability is
(Figs. 23, 24). This could be seen in Figs. 23 and 24 similar. Displacements in the model of a slope height
by the displacement contours outside the discontinuity of 200 m and k = 1, which represents the conditions
region. For k = 0.5 in Fig. 23, high values of at Pau Branco Mine has led to a maximum displace-
displacements could be observed and the contours ment of 4.2 m, close to the value reported by BVP,
suggest a circular failure in the rock material. In 3.5 m and the value of Silva’s numerical model, equal
Fig. 24, for k = 1.5, high displacements could be to 3.1 m.
observed in discontinuity region but in rock material Plastic model of a slope height equal to 200 m and
behind the slope face the displacements are zero. k = 1 has also gone into collapse, as Silva’s numerical

123
 Author's personal copy
Geotech Geol Eng

model of Pau Branco Mine. The failure mode is Diláscio, MV (2004) Estudo Computacional do Mecanismo de
similar in both models; circular failure along rock Tombamento Flexural em Filitos. Dissertação de Mestra-
do—Programa de Pós-graduação em Geotecnia—Univer-
material and buckling along discontinuities. sidade Federal de Ouro Preto, Ouro Preto, p 117
The region of QF has several geotechnical prob- BVP Engenharia (2007) Parecer Geotécnico da Inspeção dos
lems in phyllite slopes and there is a lack of input data Taludes da Cava, Mina Pau Branco Vallourec and Man-
on its properties. This work contributed to the nesman (internal report), p 24
Froldi P, Lunardi P (1995) Buckling failure phenomena and
generation of phyllite input data for further stability their analysis. In: Rossmanith HP (ed) Mechanics of join-
analysis. ted and faulted rock. Balkema, Rotterdam, pp 595–604
Remediation measures, such as slope angle optimi- Giani GP (1992) Rock slope stability analysis. Balkema, Rot-
zation, the pushback implementation effects and so on, terdam, p 361
Lopes MC (2006) Comportamento geotécnico e mecanismo de
for open pit overall slope stability can be undertaken ruptura em rochas brandas—Mineração Córrego do Sı́tio.
using the results of this study. Dissertação de Mestrado, Programa de Pós-graduação em
Engenharia Mineral—Universidade Federal de Ouro Preto,
Ouro Preto, p 94
Lopes MC, Lana MS, Ribeiro LFM, Silva CHC (2007) Com-
portamento geotécnico de rochas brandas na mina Córrego
References do Sı́tio, Santa Bárbara (MG). Revista Escola de Minas
(REM), Ouro Preto 60(3):471–476
Adhikary DP, Mühlhaus HB, Dyskin AV (2001) A numerical Rocscience Inc., (2008) Phase version 7.0—finite element
study of flexural buckling of foliated rock slopes. Int J analysis for excavations and slopes. [Link],
Numer Anal Meth Geomech 25:871–884 Toronto, ON, Canada
Bandis SC, Lumsden AC, Barton NR (1983) Fundamentals of Silva CHC (2010) Comportamento geotécnico dos filitos do
rock joint deformation. Int J Rock Mech Min Sci Geomech talude oeste da Mina Pau Branco. Dissertação de Mestrado,
Abstr 20(6):249–268 Programa de Pós-graduação em Engenharia Mineral—
Brady BHG, Brown ET (2004) Rock mechanics for under- Universidade Federal de Ouro Preto, Ouro Preto, p 111
ground mining. Kluwer, Dodrecht, p 159 Sjöberg J (1999) Analysis of large scale rock slopes. Doctoral
Cavers DS (1981) Simple methods to analyze buckling of rock thesis, Lulea University of Technology, Lulea, Sweden,
slopes. Rock Mech 14:87–104 p 788
Cosserat E, Cossert F (1909) Theorie des corp deformables.
Hermann, Paris

123

View publication stats