English
Italiano
Forgot your password?
New user?
Document Actions
ijege-13_bs-bozzano-et-alii.pdf
by different external factors (e.g. rainfalls, tunneling,
etc) made it possible to better understand the dynam-
ics of the slope-infrastructure system and to refine
the numerical model of the slope by using the finite
difference code FLAC 7.0. Such a numerical model
was implemented by applying a continuum equivalent
approach to the involved jointed rock mass, which
was considered as a visco-plastic material in order to
account for the time dependent behavior. At present
promising results have been obtained, especially in
terms of assessment of stress-strain variations due to
external forces (both environmental and man-induced)
and, thus, of forecasting the activation/reactivation of
slope instabilities.
gy model of a slope is a concept well established in
the scientific and technical community, especially if
facing on large infrastructures (see the “milestones”
such as the observational approach proposed by T
scientific and technical community facing on large in-
frastructures. The engineering-geology model is in fact
a fundamental informative layer to understand and pre-
dict the structure-slope interactions and to design stabi-
lization countermeasures. Such an issue has a relevant
role in the case of unstable slopes: at this regard the
Vajont case history represents a worldwide reference.
harmonization of engineering-geology models and
monitoring data can be achieved by the implementa-
tion of stress-strain numerical models, that represent
a validating tool for the engineering-geology models,
by collecting the monitoring data and by refining, via
calibration analyses, the rheological behaviors, i.e.
the stress-strain constitutive laws. In this frame, our
experience is referred to an unstable slope involved
in a tunnel excavation. A very detailed engineering-
geology model was built by means of several in situ
and laboratory investigations. The availability of
monitoring data with high temporal and spatial reso-
surfaces or zones and of definition of state of activity
– can then confirm the preliminary conceptual engi-
neering-geology model or modify it according to the
collected data. Once the engineering-geology model is
well constrained, it can be transferred into a numerical
model, which is the most powerful tool for backward
and forward analyses: in a calibration phase, the corre-
spondence between the numerical simulations and the
results of monitoring (then the actual displacements/
deformations) becomes the best criterion for assess-
ing the goodness of the reference engineering-geology
model, which can be eventually modified in a sort of
iterative process until the achievement of a final, sat-
isfactory model.
data, i.e.: actual displacements measured by monitor-
ing, and the results of stress-strain modeling must find
their reciprocal consistency. Once the engineering-
geology model and, then, the numerical simulation of
stress-strain relations is appropriately tested in accord-
ance with the criteria above described, the numerical
modeling becomes a tool of fundamental importance
for the prediction of deformations expected given the
changes of stress state induced – for example – in
stabilization countermeasures. Such an issue acquires
greater importance in the case of unstable slopes: in
this regard, the dramatic effects caused by the occur-
rence of the Vajont landslide on 1963 due to the dam
construction and its rapid drawdown still represents a
strong lesson about the importance of taking into ac-
count interactions between slopes and infrastructure
both in the design phase and in the executive one.
methods and techniques of investigation (both on site
and in laboratory) able to depict a detailed cognitive
framework for several purposes within the geo-haz-
ard related fields (e.g.: b
be investigated as well as on the number of investi-
gations compatible with the dedicated budget, a more
or less significant level of uncertainty still persists. In
the case of unstable slopes, a fundamental help comes
from monitoring activities, which at present benefit
from a wide variety of techniques, both traditional
and innovative, able to measure at different scales,
depths and resolutions the actual displacements and/
or the external acting forces (i.e., rainfall, dynamic in-
an unstable slope (Fig. 1). In this context during the
start-up works for the realization of a tunnel entrance,
a shallow translational landslide (with a volume of
about 10
event, the Research Centre for Geological Risks
(CERI) of the University of Rome “Sapienza” car-
ried out detailed engineering-geological investiga-
tions and surveys (geomorphological, geological
and geomechanical field surveys, boreholes, seismic
surveys and laboratory tests of samples) on the slope
in order to define a reference model to explain the
occurrence of the landslide.
Pleistocene sandy, marine deposits, while few me-
ters of sandy colluviums cover the slope. Moreover,
geological and geomorphological evidence of an old
deep roto-translational slide with a total volume of
about 1
also recognized (Figs. 1 and 2). The main sliding sur-
face (up to 50 meters deep) of the old landslide and
many secondary sliding surfaces were reconstructed
by geomorphological surveys and stratigraphic logs
(Fig. 2). Minor shallow translational movements of
about 1
can include the landslide event which destroyed part
of the already-built excavation structures.
tion works. According to this model, the reactivation
of both the whole old landslide and part of it are
plausible because of slope re-profiling and excava-
tion activities. Therefore, some stabilization work
was performed and a continuous monitoring system
was used to: i) monitor the evolution of the slope
under undisturbed conditions and during construc-
tion; ii) predict the occurrence of critical conditions,
if any. This system would also optimize planning,
design and construction activities and protect the
workers during the tunnel construction (b
the monitoring returns to be of paramount importance
to check, in the process, the adequacy of the predic-
tions made in terms of deformation/displacement. At
the end of this process it could be possible to identify
critical thresholds of significant monitored values that
can anticipate paroxysmal phases in the slope evolu-
tion, thus making the coupling of numerical models
and monitoring data a unique forecasting and plan-
ning tool that can contribute to undertake efficient risk
reduction policies, such as early-warning procedures.
The recent management of the landslide in Preonzo
- Switzerland (l
of conceptual modeling based on validation of the en-
gineering-geology model by the integration of moni-
toring and numerical modeling. The here presented
case-study is referred to an unstable slope involved
in a tunnel excavation. A very detailed engineering-
geology model was built by means of several in situ
and laboratory investigations. In addition, the slope
has been monitored for five years by means of an in-
tegrated monitoring platform made up of “punctual”
and areal displacement monitoring devices, integrated
with piezometers and rain-gauges. The availability of
data referred to slope instability episodes triggered
by different external factors (e.g. rainfalls, tunneling,
etc.) made it possible to better understand the dynam-
ics of the slope-infrastructure system and to refine the
numerical model of the slope.
IDS S.p.A.), a weather station and an automatic
photo camera (from November 2007) (b
The integrated remote platform was hosted in a
posite to the landslide one, at a distance ranging from
700 to 900 m (b
with a sampling rate of about 5 minutes. Up to now,
more than 16,000 photos, 250,000 measurements of
weather data and 250,000 SAR maps have been col-
lected. Inclinometric monitoring of the slope is manu-
ally performed since June 2007 with a sampling rate
ranging from 15 days to few days depending on the
working phases. The ‘Casagrande’ piezometers in-
stalled at the depth of 48 m below the ground level
were monitored with the same frequency of the incli-
nometers but they never recorded presence of water.
Topographic and load cells monitoring of man-made
structures (anchored bulkheads) is active since Sep-
tember 2008 with a sampling rate ranging from 15
profile of the slope. Furthermore, three anchored
bulkheads made of 0.1 m diameter and 10-22 m
long piles were realised. Bulkheads were coupled
with the deeper part of the slope by 35 m long cable
inclined of 35° from the horizontal plane. Anchors
were cemented in the last 12 m and the three bulk-
heads were realised from August 2008 to January
2009 moving downslope (Fig. 3 and monitoring data
in Fig. 4). The tunnel excavation started in Novem-
ber 2009 and was stopped in February 2010 (Fig.
5). A total amount of 28 m of tunnel were excavated
following three main phases before being interrupt-
ed due to the high value of displacement recorded
on the slope (Fig. 5). The slope was continuously
monitored starting from November 2007. The over-
all monitoring system was planned and realized in
order to record both shallow and deep displacement
of the natural slope and man-made structures. The
monitoring system consists of:
GPa and 34.04-35.04 GPa for the bedrock and for the
landslide rock mass, respectively. According to the
ISRM standard procedures (ISRM, 2007), the indexes
Ib (block size index) and Jv (number of joint per cubic
metre) were measured by geomechanical scanlines on
the outcropping rock masses. An average Jv value of
about 19 joints for cubic meter and a Ib of about 11 cm
resulted for the bedrock (MRC in Fig. 2), correspond-
ing to the rock mass class H of Table 1, while a Jv of
about 21 joints for cubic meters and Ib ranging from
8.1 to 9.6 cm resulted for the landslide mass (RL and
DRL in Fig. 2), corresponding to the rock mass classes
L and M of Table 1. Moreover a Q rock mass class
was attributed to the landslide surface (Ls) which is
characterised by an average Jv of 30 joints for cubic
meters and an average Ib of 6.3 cm. By using the two
ISRM geomechanical indexes Jv and Ib, a Rock Index
(RI) was determined, which is based on a statistical
algorithm (M
144 geomechanical scanlines which were performed
on the rock masses outcropping within an area of 100
km
tion of the data series of Jv and Ib in series of data with
mean equal to zero and variance equal to one by the
relation (M
the standard deviation of the data x
rock mass classes by using the RI index.
The convergence monitoring of tunnel had a sampling
frequency ranging from a day to a week, with a step of
7 m for depth of excavation.
ity of displacement on the order of 0.05 mm/h. Immedi-
ately after the beginning of the excavation the velocity
of bulkheads suddenly increased reaching maximum
values of 0.75 mm/h, with acceleration and decelera-
tion peaks on the order of 0.02 mm/h
a maximum displacement of about 100 mm was record-
ed on the first order of bulkhead. In the last two phases,
the interferometric monitoring allowed us to clearly
recognize a typical creep behaviour.
stratigraphical logs and CID triaxial tests. The geome-
chanical characterization of the old deep roto-trans-
lational slide was achieved by data coming from the
stratigraphic logs (Fig.1). The CID triaxial tests were
performed on some samples of gneiss at the “Labora-
torio della Provincia Autonoma di Trento” geotechni-
cal laboratory to evaluate both strength and deforma-
tional parameters of the intact rock (M
rock from triaxial laboratory tests vary in the range of
50–60 GPa and the UCS resulted about 94 MPa (b
they were derived for each rock mass class after a
stress-strain calibration process via numerical mod-
eling, by assuming a creep rheology according to the
visco-plastic Burgers model.
slope, by the finite difference code FLAC 7.0 (ITAS-
CA, 2011). Geological features as well as structural
elements (such as anchored bulkheads, retaining
walls) were also modelled in order to follow the
stress-strain changes induced within the slope by the
on-going human activities. The rheological behaviour
of the involved rock mass was reproduced according
to an equivalent continuum approach (s
rameter values the intact rock mass properties as well
as the jointing conditions (Hoek and Brown failure
criterion combined to Mohr-Columb failure criterion).
RL and Ls lithotechnical units (Fig. 2). Respectively:
1) a Burgers visco-plastic model was assumed for the
Tab. 1 summarises the parameter values attrib-
b
Burger model were selected in the range 1019-1023
Pa•s and calibrated to match morphological evidenc-
es and the displacements recorded as a consequence
of the landslide.
the Kelvin-Voight visco-elastic element were always
assumed to be one order of magnitude higher than
the ones used for the visco-plastic Maxwell element.
Based on the performed numerical calibration, the
here considered viscosity values result the lowest for
justifying the landslide activation in the present slope
shape with respect to the river valley evolution. On
the other hand, for calibrating the viscosity values for
RL, DRL and Ls a best fit was performed between
the monitored displacements, referred to the different
excavation and re-shaping steps within the landslide
mass and the numerical modeled ones.
performed for calibrating the rock mass rheological
behavior of the landslide mass. A slope re-shape was
realized starting from March 2007 and three orders
monitored displacements due to the re-shaping of the
considered slope and the realization of reinforcements
(bulkheads). During the here considered time interval
the slope experienced a total amount of displacement
of more than 100 mm. By this back-simulation the
best rheological model describing the rock mass me-
chanical behavior was determined and its parameters
were calibrated. The stress-strain effects related to the
slope re-shaping and reinforcement were referred to
the engineering geology model and it can now pro-
vide scenarios of possible effects due to destabilizing
actions external to the slope as well as to the manage-
ment of the road tunnel.
rock by applying the continuum equivalent appro-
ach to the time dependent behavior. Up to now, such
an approach has been experienced and validated by
other authors (r
demonstrated the reliability of such an approach also
for the visco-plastic behavior applied to sliding slopes
emphasized by coupling it with time and space high
resolution monitoring data.
slope after the cut of corresponding parts of the slope.
Horizontal displacements up to some centimeters in-
volved the landslide mass after the occurrence of the
March 2007 landslide (Fig. 6a); further centimeter-
scale displacements added to the previously occurred
ones as a consequence of the re-shaping and of the cut
of the trenches (Fig. 6b and 6c).
sulting by the numerical model after the rheological
calibration (i.e. according to the rock mass parameters
summarized in Tab. 1); the numerically computed
displacements are compared with the ones measured
along the inclinometer borehole and at the anchored
bulkheads during the re-shape and consolidation
works. The comparison demonstrates the very good
fit among the modeled and the measured displacement
values, thus providing a stress-strain scenario which
strongly reproduces the actual behavior of the slope,
observed during the consolidation works.
toring and numerical modelling in the achievement of
a conceptual model based on a engineering-geology
one with reference to a specific case-study. In this fra-
me, the coupling of numerical models and monitoring
data could represent an effective forecasting and plan-
ning tool that can contribute to undertake efficient risk
reduction policies.
The issue 16 volume 01 has been published