Document Actions

ijege-13_bs-croce-et-alii.pdf

background image
Italian Journal of Engineering Geology and Environment - Book Series (6) www.ijege.uniroma1.it © 2013 Sapienza Università
Editrice
399
DOI: 10.4408/IJEGE.2013-06.B-38
GEOTECHNICAL ANALYSIS OF AN ANCIENT DAM BUILT
IN UNFAVOURABLE GEOLOGICAL CONDITIONS
P
aolo
CROCE, G
iusePPe
MODONI, a
lessandro
RASULO & M
ichele
SAROLI
University of Cassino and Southern Lazio - Cassino, Italy
tail. The available observations have been evaluated in
order to gain an over all interpretation of the reservoir
behaviour, with particular reference to the moraine de-
posit. Stability analyses have finally been performed to
check the safety level of the reservoir.
K
ey
words
: dam, moraine, seepage, displacements, monito-
ring, back-analysis
INTRODUCTION
At the beginning of the twentieth century, several
dams were built in the Italian Alps with the aim of
providing water reservoirs for electrical supply. Most
of these dams are still on service, even though they
were designed and constructed following out of date
concepts and techniques. Moreover, the locations cho-
sen for such dams would often be questionable, from
the geological and geo-morphological stand point, ac-
cording to present safety criteria. The overall safety of
such dams and reservoirs should then be assessed by
means of modern investigation methods.
One of such cases is the Alpe Cavalli dam,
built in the 1920’s, in the north-western Italian Alps
(Fig. 1). The dam is located in a very peculiar geo-
morphological environment. In particular, while the
right abutment is located on a moraine deposit the
dam foundations are placed on the fluvio-glacial de-
posit and the left abutment rests on a cliff of meta-
morphic rocks.1
The reservoir has been effectively operating for
ABSTRACT
The paper reports on the behaviour of the Alpe
Cavalli dam, built in the Italian Alps between 1922
and 1926 and still in operation for electric power sup-
ply. The dam, which is 31 m high, made by dry ma-
sonry, is waterproofed by a multi-layered impervious
facing, a concrete cut-off and a grout curtain in the
foundation soils. The geological conditions of the dam
site are quite peculiar since the reservoir location co-
incides with an ancient natural lake of glacial origin
which was emptied, probably after the last glaciation,
when the front moraine was eroded by the emissary
river and a deep narrow cut was opened on the left
side of the valley. For such reason the right shoulder
of the dam rests on the moraine deposit, its central
body is placed on the river bed alluvium while its left
shoulder stands on a metamorphic formation com-
posed of highly fractured rocks.
A comprehensive investigation has been conduct-
ed in order to understand the overall behaviour of the
reservoir, after about ninety years of service. For this
purpose, historical documents have first been reviewed
for reconstructing the design assumptions and the
building operations. Experimental investigations of the
moraine deposit have then been conducted, in order to
estimate the subsoil properties. Seepage flow rates and
reservoir impoundment levels have been thoroughly
analyzed, showing a strong correlation between these
two variables. The displacements of the dam body and
of the moraine deposit have also been examined in de-
background image
P. CROCE, G. MODONI, A. RASULO & M. SAROLI
400
International Conference Vajont 1963-2013. Thoughts and analyses after 50 years since the catastrophic landslide Padua, Italy - 8-10 October 2013
GEOLOGICAL SETTING
The Loranco stream, left tributary of Ovesca
river, is located in the north-western Alpine sector.
Its evolution is conditioned by morphoclimatic and
structural features, developed in a fluvial-glacial en-
vironment. A sharp change of direction in its course
(from ESE-WNW to NW-SE) is due to the different
erosion rate among the outcropped bedrocks. Gla-
cial proofs have been recognized in lateral and front
moreins, partially covered by talus and localised
on the bottom of the main valley. Moreover hang-
ing valleys are present on the left side. Because of a
transition from glacial to fluvial condition, the front
morein has obstructed the river valley thus creating
a paleolake. The fluvio-glacial deposit, interbedded
with talus, fills all the valley and lies on the meta-
morphic rockmass. Paleozoic ortogneiss and gran-
ite of “Monte Leone” are overlapped on mesozoic
schists, belonging to the piedmontese “Pietre Verdi”
series, by means of a regional thrust (S
CHMID
, 2004).
Surveys on “Pietre Verdi” have revealed schistosity
planes dipping approximately 335.50 nearby the left
shoulder. Perpendicular to the stratification,another
schistosity set has been recognized with a 110.35
trend. The “Pietre Verdi” series is characterized by
“prasinitic”, amphibolitic and calcareous schists,
permeable for cracking. From an hydrogeological
point of view, the circulation scheme of the area can
be divided in two different paths respectively for the
methamorfic complex and the morain deposit. In par-
ticular, the discontinuities of the metamorphic layers
are filled and covered by eluvial clays, whereas the
glacial-karst process is barely visible in the calcare-
about ninety years so far. However, since its first
impoundment, relevant seepage has been observed
downstream and large water flows have been continu-
ously recorded from then on, in spite of several wa-
terproofing treatments. Significant movements have
also been observed on the dam and on the moraine
abutment producing localized deformations on the
downstream face of the dam.
A thorough assessment of the safety conditions
of this old dam has already been attempted (C
roce
& M
odoni
, 2008a; 2008b). However, the critical
geological factors deserve some further explanation
which is provided in the present paper. For this pur-
pose geological and geotechnical investigations have
been performed, in order to clarify the subsoil con-
ditions. Systematic topographical measurements are
also available but only for few years.
Moreover a large amount of observations, cover-
ing a time period of about twenty years, has been col-
lected and analysed. These data concern water level
fluctuations in the reservoir, water losses through the
moraine deposit, minimum and maximum daily tem-
peratures at different depth in the reservoir and in the
atmosphere, rainfall and ice thickness.
In the following, a brief description of the reser-
voir features and of the dam history is provided first.
Results of the recent geotechnical investigations and
monitoring are presented next.
The seepage data are then analyzed and the flow
regime is simulated by numerical calculations to-
gether with the cyclic displacements of the moraine
deposit, induced by the water level fluctuations in the
reservoir. This analysis is aimed to provide a logical
explanation of the dam movements and of the ob-
served cracks on the downstream abutment.
Fig. 2 - Plan view of the Alpe Cavalli dam with the loca-
tions of the intercepted water springs (W1, W2 and
W3) and of the continuous borings (S1, S2 and
S4); the Loranco River is indicated by the thick
continuous line and the trace of the analysed cross
section through the morain deposit is indicated by
the dotted line
Fig. 1 - Partial overview of western Alpine sector between
Domodossola and Varese (A), Alpe Cavalli reser-
voir (B), topographic sketch of Alpe Cavalli lake,
scale 1:50.000 (C)
background image
GEOTECHNICAL ANALYSIS OF AN ANCIENT DAM BUILT IN UNFAVOURABLE GEOLOGICAL CONDITIONS
Italian Journal of Engineering Geology and Environment - Book Series (6) www.ijege.uniroma1.it © 2013 Sapienza Università
Editrice
401
of the right abutment formed by the moraine deposit.
However, the dam was finally located at the bot-
tom of this valley resting its right shoulder on the mo-
raine, its central body on the river bed fluvio-glacial
deposit and its left shoulder on the metamorphic rocks
(Fig. 4). The works started in 1922 and the dam was
completed in 1926. During the construction of the
dam, further doubts arose on the soundness of the rock
formation, located below the left shoulder of the dam,
which appeared highly fractured and affected by the
presence of cavities.
Concerning the river bed alluvial deposit, ten years
after the completion of the dam, an extensive campaign
of ground improvement was undertaken, consisting of
several injections of grout. Each injection, which pro-
truded into the lower rock stratum, was performed with
a 30 bars pressure in order to fill the alluvial deposit.
This measure proved to be effective since it resulted in
a water flow reduction in the drainage tubes of about
60 %. A similar subsoil treatment was also performed
in order to reduce the water loss in the fissured rock
mass of the left shoulder. However, although a reduc-
tion was obtained, still significant water losses were
observed (G
IGNOUX
& B
ARBIER,
1955).
In fact, the most difficult problem detected dur-
ing the first years of reservoir operation was the seep-
age observed through the moraine deposit of the right
shoulder, which raised serious concern also on its sta-
bility and thus on the overall safety of the reservoir.
This moraine deposit may be considered as the natural
extension of the dam. In particular, as written in one of
the original design report, even before the construction
of the dam, several water springs had been observed at
the downstream foot of the moraine deposit and local
debris slides had also been recorded downhill. After
the construction of the dam, the water flows became
ous schists. Weathered rocks underlying the morain
deposit are fairly impermeable thus representing an
acquiclude for the morainic acquifer
DAM FEATURES
The Alpe Cavalli reservoir, with a full capacity
of 8,35 millions of cubic meters, is located in the
basin of the previously mentioned paleo- lake, 1500
meters above see level (Fig. 2). The reservoir feeds
the ground water seepage mainly through the morain
body. The dam height, computed from the bottom of
its foundation is 41.60 m, and the crest is 165 meters
long. The barrage is made of masonry and is water-
proofed, on its up-hill side, by a composite multi-
layered facing of cement mortar, reinforced concrete,
bitumen and bricks (Fig. 3).
The dam foundation is made of a continuous ma-
sonry layer, about 1.30 meters thick, cemented by hy-
draulic mortar and is provided with a concrete vertical
cut-off reaching the maximum depth of 10 meters.
Since the early design studies, performed before
the first world war, the engineers were puzzled by the
large extension and the depth of the fluvio-glacial de-
posit and by its hydro-mechanical behaviour, with par-
ticular reference to the stability and the water-tightness
of the foundations and of the right shoulder of the dam.
In particular, the first design question concerned the
extension and the depth of the alluvial deposit located
below the embankment. The second one was related
to the waterproofing capacity and the overall stability
Fig. 4 - Longitudinal geo-morphological pattern at the
Alpe Cavalli dam
Fig. 3 - Cross section (a) and front view (b) of the masonry
dam (ANIDEL, 1953)
background image
P. CROCE, G. MODONI, A. RASULO & M. SAROLI
402
International Conference Vajont 1963-2013. Thoughts and analyses after 50 years since the catastrophic landslide Padua, Italy - 8-10 October 2013
ervoir. Visual inspection of the samples retrieved from
all the borings showed an alternation of dark-red gravel
layers immersed in a sandy silt to silty-gravelly sand
matrix. Undisturbed samples could not be retrieved but
several remolded samples were recovered by the three
continuous borings at different depths, in order to de-
tect their grain size distribution. Typical results of these
analysis are summarized in Fig. 6 where the soil com-
position observed at boring S2 is reported as a function
of depth. The soil can be broadly classified as a well
graded coarse grained material, being gravel and sand
the predominant fractions, with a low but not negligible
portion of fine grained soils made of silts and clays.
Although it is not shown from the grain size analy-
sis the presence of large blocks in the moraine was also
detected. However, considering the large extension of
the morain deposit, the soil can be considered to be
practically homogeneous.
Le Franc tests were also performed at different
depth along the bore holes. The permeability coef-
ficients obtained from such tests are summarized in
Fig. 7, together with a broad estimation of the Darcy’s
quite large and three most significant springs were col-
lected to be pumped back into the reservoir (W1, W2
and W3; Fig. 2).
During all the operation period significant move-
ments of the dam body have also been recorded and
vertical cracks were observed on the downhill abut-
ment, close to the left shoulder (Fig. 5). This last obser-
vation suggested to monitor the dam movement with
systematic topographical surveys.
GEOTECHNICAL INVESTIGATIONS
Due to the lack of modern tools, geotechnical in-
vestigations of the moraine deposit could not be per-
formed at the time of construction and thus the exten-
sion and the mechanical characteristics of the deposit
were largely unknown. Three continuous boring were
recently performed in different locations (Fig. 2)
A 60 meters deep borehole was first performed
starting from 1475 m a.s.l. close to the downhill toe of
the embankment. A second borehole, 80 meters deep
was then accomplished in the middle of the moraine
deposit, at 1506 n meter a.s.l.. The last borehole, 40
meter deep, was performed at 1447 meters a.s.l., near
the toe of the moraine on the opposite side of the res-
Fig. 5 - Down stream abutment of Alpe Cavalli dam with
indication of the observed vertical crack
Fig. 6 - Grain size composition of the moraine at boring S2
Fig. 7 - Permeability coefficients of the moraine
Fig. 8 - Young moduli obtained from pressumeter tests
background image
GEOTECHNICAL ANALYSIS OF AN ANCIENT DAM BUILT IN UNFAVOURABLE GEOLOGICAL CONDITIONS
Italian Journal of Engineering Geology and Environment - Book Series (6) www.ijege.uniroma1.it © 2013 Sapienza Università
Editrice
403
particular, the water emerging from three of the most
plentiful springs located at the toe of the moraine (W1,
W2 and W3 in Fig. 1) was collected and pumped back
into the reservoir.
Such a procedure allowed for a continuous record-
ing of the water flow trend through the moraine during
the period 1980-1998 As an example, the collection
of daily measures for the period 1982-1984, together
with the simultaneous measures of water level in the
reservoir, is reported in Fig. 9. A seasonal cyclic trend
can be observed from both series of data, with mini-
mum values occurring during spring and maximum
values in autumn. Such a trend is fully consistent with
the hydrological features of the Italian Alps, where
snow typically falls during winter. The time offset be-
tween the water reservoir level and the snow fall can
be explained by considering the reservoir as mainly
fed by the water coming from spring melting and
transported by the Loranco River. The level in the res-
ervoir is the result of the balance between the amount
of water coming from the Rio Loranco and the one
discharged for electric supply.
The similitude of the two curves of Fig. 8a is the
most clear sign of the correlation existing between
reservoir level and seepage through the moraine. A
quantitative evaluation of the time delay existing
between these two quantities has been obtained by
relative shifting of the two curves reported in Fig. 9a
by a period T and evaluating the corresponding cross
correlation factor defined in Fig. 9b. The results of
such analysis clearly show that the best superposi-
tion of the two curves is obtained with a practically
nil delay, accordingly with the high permeability of
the moraine.
The total amount of water collected at the three
springs (Q) and the reservoir level H are directly com-
pared in Fig. 10, scaling both of them by their maxi-
coefficient provided by the Hazen correlation based on
grain size distribution (h
azen
, 1911). Generally a scat-
ter of three orders of magnitude can be observed from
all the reported data but, when looking only at the Le
Franc tests, the permeability coefficient ranges from
2x10
-4
to 3x10
-3
cm/s, which is typical of a fine sand.
With regard to the mechanical properties, it is re-
called that the deformation and strength characteristics
of fluvio-glacial soils are generally very difficult to
evaluate. In fact undisturbed sampling is not possible
due to the large size of the soil particles. Furthermore
soil is very heterogeneous at the scale of typical situ
tests, due to the occurrence of large blocks and thus the
meaning of these tests results is very low.
For such reasons the compressibility of glacial
origin soils has been evaluated by C
ROCE
et alii. (1963)
back analyzing the settlements measured at the foun-
dation of three dams built in the Italian Alps after the
Second World War. In the hypothesis of linear elastic
response, a Young modulus ranging between 30 and
160 MPa was obtained from such study.
In the present case, a rough estimate of the Young’s
modulus is provided by the results of several pres-
suremeter tests (M
ENARD,
1975), performed along the
three boreholes at different depths.
As reported in Fig. 8, the values of the Young mod-
uli E calculated from the pressuremeter tests are quite
variable, ranging from a minimum of 21 to a maximum
of 129 MPa and there is not a clear dependence on the
depth. The observed values are however consistent with
the ones previously quoted from the literature.
SEEPAGE THROUGH THE MORAINE
Soon after the first impoundment of the reser-
voir, significant water losses were observed and then
monitored through both the dam embankment and the
moraine body, being these latter the most relevant. In
F
ig. 9 - Reservoir level and water loss in the period 1982-1984 (a) (Qave=0.165 m3/s ; Zave=1488.4 m a.s.l.) and cross correla-
tion among the two series (b)
background image
P. CROCE, G. MODONI, A. RASULO & M. SAROLI
404
International Conference Vajont 1963-2013. Thoughts and analyses after 50 years since the catastrophic landslide Padua, Italy - 8-10 October 2013
mum observed value. The plot confirm the strong cor-
relation between the two variables. However it is also
seen that there is a non negligible amount of flow rate
(more than 10% of the maximum value), not directly
related to the water level in the reservoir. It is inferred
that such portion of water comes from natural flows
through deeper water bearing strata.
One final important question regards the stability
of such correlation along with time, which could be
a sign of potential erosion through the moraine. It is
easy to argue from Fig. 11, where the water flows
corresponding to fixed impounding levels are report-
ed as a function of time, that the average flow rate is
almost constant throughout the years, thus confirm-
ing that the permeability of the moraine mass is sub-
stantially constant.
DAM AND MORAINE DISPLACEMENTS
The dam movements have been monitored by a
multiple topographical survey system (Fig. 12). Align-
ment measurements are available since the beginning
of 1998 for two points of the dam body (A’ and C’)
from two stations (A and C) located on the metamor-
phic rock at the left of the dam. Triangulation measure-
ments are available from the period 2001-2002 for one
target on the moraine, taken from a station located on
the metamorphic rock (B).
The comparison between the displacements re-
corded on alignments A-A’ and C-C’ (Fig. 13) provides
basic information on the deformation mechanisms of
the dam. In fact, while the position of point C’ located
close to the right dam shoulder is almost stable (the
recorded displacements are less than 1 mm), the point
A’ located at the centre of the dam shows every year
cyclic movements with a double amplitude equal to
about 6 mm. In the long period this latter movements
provide also a permanent displacement increasing
with an average rate of 1.6 mm/year. It can be inferred
that this trend is strictly related to the response of the
moraine deposit on the right dam shoulder to the cy-
clic loading produced by the water level excursion in
the reservoir.
The horizontal displacements recorded at point B
of the moraine in the two years 2001 and 2002 have
been composed on the alignment orthogonal to the mo-
raine crest (Fig. 12) and related to the water level in the
reservoir (Fig. 14).
Even these measurements show the close relation
existing between the deformation in the moraine body
and the reservoir level. In particular the observed dou-
ble amplitude of the cyclic displacement at the top of
the moraine is equal to 13 mm. In such short measure-
ment period the accumulated displacements are rather
low and thus the moraine stress strain behavior is al-
most reversible.
Fig. 11 - Progression with time of water losses measured at
three different reservoir levels
Fig. 12 - Topographical survey system installed at Alpe
Cavalli
Fig. 10 - Water loss versus reservoir level in the Alpe Cav-
alli moraine (period of observation 1980-1998)
background image
GEOTECHNICAL ANALYSIS OF AN ANCIENT DAM BUILT IN UNFAVOURABLE GEOLOGICAL CONDITIONS
Italian Journal of Engineering Geology and Environment - Book Series (6) www.ijege.uniroma1.it © 2013 Sapienza Università
Editrice
405
tests (Fig. 7). This analysis has been performed for
different reservoir levels and its results are compared
with the experimental ones (previously reported in
Fig. 10) in the non dimensional plot of Fig. 16. In this
plot, the measured and computed flow rates have been
divided by their maximum values and the water level
has been scaled by the minimum and maximum levels
in the reservoir. The different shape of the theoretical
and experimental curves can be explained considering
that the assumed two-dimensional scheme may not be
fully representative of the entire moraine deposit and
NUMERICAL ANALYSIS
A numerical simulation of the observed phenom-
ena has been performed by means of a finite elements
code (P
LAXIS
vers. 7.0, 1998) in order to analyse the
response of the moraine with regard to seepage and
the stress-strain behaviour of the moraine deposit
A two-dimensional model (Fig. 15) has been im-
plemented at the cross section reported by the dotted
line in Fig. 2 (i.e. close to the point B’, where meas-
urements of displacement were available). In the
model, the moraine body is superimposed to a thick
layer of metamorphic rocks whose position has been
estimated based on the results of the subsoil investiga-
tions supported by geological considerations.
The seepage conditions through the moraine have
been initially simulated by supposing the whole mo-
raine body as having a unique constant permeability
coefficient (k=1.16*10-3 cm/s) equal to the mean val-
ue of the data obtained by the results of permeability
Fig. 15 - Two-dimensional model of the moraine at cross
section reported by dotted line in Fig. 2
Fig. 16 - Simulation of seepage through the moraine deposit
Fig. 13 - Alignment measurements recorded on A-A’ and C-C
Fig. 14 - Recorded horizontal displacements at point B (a)
and reservoir impounding level (b)
background image
P. CROCE, G. MODONI, A. RASULO & M. SAROLI
406
International Conference Vajont 1963-2013. Thoughts and analyses after 50 years since the catastrophic landslide Padua, Italy - 8-10 October 2013
curvature and a rightward bending compared to the
polynomial fitting curve of the experimental data. It
is worth observing that such divergence with meas-
ured data would be more pronounced if plasticity
were introduced in the soil model and that a non lin-
ear elastic model with increasing stiffness would be
more appropriate to fit the experimental curve. Such
observation is fully consistent with the stress-strain
response observed from laboratory experiments on
gravels repeatedly subjected to large amplitude load-
ing cycles (a
nh
D
AN
& K
OSEKI
, 2004; M
odoni
et
alii, 2006). In particular, the plot of Fig. 17 clearly
shows the change of the stress-strain response due
to repeated cyclic loading, with a marked increase
of stiffness.
In the left part of the plot, the deviator stress is
also reported versus the elastic distortional strains,
these latter calculated by performing small strain
unloading-reloading cycles (T
ATSUOKA
& S
HIBUYA,
1992). It is seen that the soil response at the end of
the test (i.e. after a large number of loading cycles)
becomes similar to the elastic response, with an in-
versed convexity of the stress strain curve compared
to the initial loading. As a further consequence of
such strain accumulation, the post cycles stress-
strain response of soil becomes more brittle, but with
a noticeable increase of the soil stiffness and strength
(M
ODONI
et alii, 2006). These results suggest that cy-
clic loading produced on the moraine by the seasonal
excursion of water level in the reservoir should have
a positive effect in terms of stiffening and strength-
ening of the soil response but with the negative effect
represented by the progressive strain accumulation
(Fig. 11). This latter effect should be carefully con-
trolled by continuously monitoring the movements
of the dam and of the moraine.
that the moraine permeability may not be constant.
Concerning this latter issue, a possible variation of
permeability with depth has been considered. Four
different layers, parallel to the bedrock, have thus
been introduced in the numerical model and different
permeability coefficients have been assigned to each
of them (Fig. 15). The most appropriate values of k
(Figs. 7 and 15) have been established by a trial and
error procedure. It can be seen that by assuming these
coefficients increasing from the bottom to the top of
the moraine a better simulation of the observed ex-
perimental trend is obtained (Fig. 16).
Under these seepage conditions, a numerical anal-
ysis has then been performed on the displacements
induced by the water fluctuation in the reservoir. For
this purpose, the soil has been modelled as a linearly
elastic material and the horizontal displacements at
the top of the moraine deposit (Fig. 17a) have been
computed at several impounding levels for different
Young’s moduli. A comparison of simulated displace-
ments with those observed at point B’ (Figs. 12 and
14) is reported in Fig. 17b.
This analysis shows that the elastic modulus of
the soil model necessary to simulate the observed
displacements of the moraine is larger than those
evaluated by the pressuremeter tests (Fig. 7). Fur-
thermore, all the simulation curves show a larger
Fig. 18 - Cyclic triaxial shearing of gravel (M
odoni
et alii,
2006)
Fig. 17 - Pattern of calculated displacements (a) and com-
parison between measured and calculated hori-
zontal displacements at point B’ (b)
background image
GEOTECHNICAL ANALYSIS OF AN ANCIENT DAM BUILT IN UNFAVOURABLE GEOLOGICAL CONDITIONS
Italian Journal of Engineering Geology and Environment - Book Series (6) www.ijege.uniroma1.it © 2013 Sapienza Università
Editrice
407
this latter is partly founded on moraine and alluvial
soil, partly resting on a metamorphic rock formation.
Such difference is responsible for a significant strain
concentration on the right shoulder of the dam which
have resulted in visible cracks on the downstream
facing. The results of numerical back analysis have
also shown that the soil response is much stiffer than
estimated by pressuremeter tests and similar to a non
linear elastic medium. Such response is consistent
with laboratory observation on gravelly soils sub-
jected to a large number of loading cycles.
With regard to the stability of the moraine, a
parametric analysis has been carried out by c
roce
&
M
odoni
(2008b). Considering the experience so far
cumulated on materials having the same grain size
composition and density as those of the morainic
deposit, this analysis suggests that deep instabil-
ity mechanisms are unlikely to occur, while some
concerns may arise with respect to shallow sliding
surfaces. Such interpretation is supported by the evi-
dence that the moraine as a whole has never shown
any signs of instability. On the other hand, shal-
low mechanisms confined to the top cover of the
morainic deposit had already been observed by the
designers along the Loranco river downstream of
the dam: much lower than the "piano dei Cavalli",
to the right of the course of Loranco river ... a large
landslide named "Rovina" is taking place (N
OVARESE
& S
TELLA
, 1919). Still visible landslides have been
more recently reported (F
AITELLI,
1999) as shown in
the schematic geological map (Fig. 4).
CONCLUSIONS
The case reported in the present paper has regard-
ed a masonry dam built in the Italian Alps in the years
1922-1926. This dam has been continuously in opera-
tion for electrical purposes, and still is, even though
large seepage flows and significant dam movements
have been recorded along the years. Due to the lim-
ited knowledge of the subsoil properties and to the
lack of well established calculation methods, at the
beginning of the last century, the design was car-
ried out mostly based on empirical rules. The paper
provides an updated analysis of the present working
conditions of the dam by combining the results of
a geotechnical investigations campaign, a monitor-
ing plan and different numerical analyses. The most
peculiar aspect observed in this case is the complex
interaction between the reservoir, the dam body and
the different subsoil formations. In particular, a pre-
dominant role is played by the moraine deposit form-
ing a natural prolongation of the dam along its right
shoulder. Geological studies and geotechnical inves-
tigations have provided basic information on this de-
posit but the grain size composition of the moraine,
consisting mainly of coarse grained material, has
limited the geotechnical information which could be
retrieved by laboratory and site investigations. This
gap has then been filled by direct observations and
by complementary numerical analyses. The former
has provided a logical correlation between reservoir
level excursions, water seepage and movements of
the moraine and of the masonry dam. In particular,
REFERENCES
a
nh
d
an
L.Q. & K
oseKi
J. (2004) - Effects of large number of cyclic loading on deformation characteristics of dense granular
materials. Soils and Foundations, 44: 115-123.
ANIDEL (1953) - Le dighe di ritenuta degli impianti idroelettrici italiani. Technical Report, 31, Milano.
c
roce
a., d
olcetta
M., F
inzi
d. & M
artinelli
D. (1963) - Compressibility of soils of glacial origin. Proc. of the European Conf.
on Soil Mechanics and Foundation Engineering, Wiesbaden.
c
roce
P. & M
odoni
G. (2008a) - Analysis of dam bahavior after eigthy years of service. 6
th
Intenational Conferences on Case
Historie in Geotechnical Engineering. Arlington, VA, August 11-16, 2008.
c
roce
P. & M
odoni
G. (2008b) - Analisi geotecnica di una diga di antica costruzione. Rivista italiana di Geotrecnica, 4/2008.
F
aitelli
G. (1999) - Diga di Alpe Cavalli: Indagine geologica conclusiva.
G
iGnoux
M. & B
arBier
R. (1955) - Geologie des barrages et des amenagements hydrauliques. Paris, Masson.
h
azen
A. (1911) - Discussion on dams on sand foundations. Trans. ASCE, 73: 199.
M
enard
L. (1975) - The Menard Pressuremeter: Interpretation and Application of the Pressuremeter Test Results to Foundations
Design. Sols-Soils, 26, Paris, France.
M
odoni
G., a
nh
d
an
L.Q., K
oseKi
J.& M
aqBool
S. (2006) - Strain cumulation of gravels produced by cyclic loading. Soil stress-
strain behavior: measurement, modeling and analysis. Springer, Rome March 16-17 2006.
background image
P. CROCE, G. MODONI, A. RASULO & M. SAROLI
408
International Conference Vajont 1963-2013. Thoughts and analyses after 50 years since the catastrophic landslide Padua, Italy - 8-10 October 2013
P
laxis
vers
.7. (1998) - Finite Element Code for Soil and Rock Analysis. Balkema.
s
chMid
S.M. (2004) - Tectonic map and overall architecture of the Alpine orogen. Eclogae Geologicae Helvetiae, 97: 93-117.
t
atsuoKa
F. & s
hiBuya
S. (1992) - Deformation characteristics of soils and rocks from field and laboratory tests. Report of the
Institute of Industrial Science of the University of Tokyo, 37.
Statistics