1555
Case study of the post-earthquake behavior of a CFRD dam
Étude de cas sur le comportement post-sismique d'un barrage CFRD
Núñez E., A. Sfriso O.
University of Buenos Aires, Argentina
ABSTRACT: Potrerillos dam is a 120m high CFRD dam recently built in Mendoza, Argentina. The dam was built on top of a
riverbed containing loose sand seams. Concern about the seismic stability of the dam derived in the construction of a large
downstream buttress, not considered in the original design and bid. The key question that arose during the consequent claim is: was
the buttress essential for the dam’s stability? In the paper, a twofold analysis of the problem is performed: i) the analytical methods
employed in the original discussion are briefly introduced and discussed; ii) a simple static pushover analysis is performed
numerically, where the sand seams are forced to follow the full strain path through contractive to dilating to critical state. In both
analyses, the thrust required to displace the dam body is compared to the hydrostatic pressure to assess on dam’s stability. The results
obtained using both methods are compared. The main conclusion is that the self-weight of the dam produces an increase in the
undrained shear strength of the sandy seams that yields the downstream buttress not essential.
RÉSUMÉ: Le barrage de Potrerillos est un type CFRD de 120 m d'hautesse construit récemment à Mendoza, Argentine. Le barrage
fut construit dans un lit de rivière qui contient des strates de sable lâche. Il y avait une inquiétude sur la stabilité sismique du barrage
laquelle vient dérivée de la construction d'un contrefort en aval, qui n'était pas considérée dans le design originaire et l'offre. La
question clé qui surgit pendant la réclamation conséquente c'est : est-ce que le contrefort était essentiel pour la stabilité du barrage?
Dans l'essai, une analyse double du souci est proposée: i) les méthodes analytiques utilisés dans la discussion originaire sont
brièvement introduits et discutés; ii) une simple analyse statique faible est réalisé numériquement, où les strates de sable sont forcés à
suivre le parcours de tension complète de la contraction en passant à la dilatation jusqu'à l'état critique. Dans touts les deux analyses,
la poussée nécessaire pour déplacer le corps du barrage est comparée avec la pression hydrostatique en évaluant la stabilité du
barrage. Les résultats obtenus en utilisant les deux méthodes sont comparés. La conclusion principale c'est que le poids il-même du
barrage entraîne un augment de la résistance au coupe et non drainée des strates de sable, lequel ne fait pas nécessaire le contrefort.
KEYWORDS: CFRD dam, seismic analysis, liquefaction, numerical analysis.
1 INTRODUCTION
Potrerillos is a 120m high CFRD dam located in Mendoza,
Argentina, in a region of high seismicity (Carmona et al 2004,
Zabala et al 2006) yielding a design earthquake with a
magnitude M = 7.7 and a peak ground acceleration PGA = 1.02.
The dam has upstream and downstream slopes of 1.5:1 and
1.8:1 and is founded on an alluvial riverbed made of dense
gravels up to 70m thick. The hydraulic closure of the upstream
face in the riverbed is furbished by a conventional concrete slab
connected to a cast-in-situ diaphragm wall by a floating plinth.
The project has been thoroughly described elsewere (e.g.
Barchiesi et al 2006, Núñez 2010a). The cross section of the
dam is shown in Figure 1 (Barchiesi et al 2006).
Figure 1. Cross section of Potrerillos dam (Barchiesi et al 2006).
Geotechnical campaigns performed before the construction of
the dam failed to identify sand and low plasticity silt lenses, 2m
to 4m thick, that are located some 30m below the mean
foundation level and that were sistematically found during the
execution of the diaphragm wall and the installation of piezo-
meters. It could not be determined and is yet uncertain if these
sands and silts lie in isolated pockets or form a continuous layer
throughout the dam foundation; in the latter case, they might
pose a significative risk of failure due to liquefaction / strength
loss during and after a major earthquake.
While client engineers assumed that the layer was continuous
and liquefiable, design and construction engineers did not
believe the continuous layer assumption and argued that the
existence of those materials was of no concern. As the
construction of the dam was highly advanced and the found-
ation surface was no longer accessible, only a few additional
SPT tests were scheduled and performed, where low plasticity
silts (4 < PI < 8) and medium dense sands were found again.
Client engineers sat back on their worst case scenario of a
continuous layer, assumed an undrained shear strength of 25kPa
to 50kPa for the layer and concluded that the dam would not be
stable during earthquake loading.
Based on this body of information, the client instructed the
contractor to build a 19m high downstream buttress embank-
ment to improve the overall seismic stability of the dam. The
contractor built the buttress and started a claim that soon
escalated to the Supreme Court of Justice of Mendoza. The first
author was nominated as expert witness and was asked only
one, yet crucial question: was the buttress indispensable for the
stability of the dam?
To answer it, he performed an independent study of the case
based on a thorough review of the geotechnical data and a few
rigid body stability calculations. This paper summarizes such
study, presents its main conclusions, and briefs the results of an
alternative numerical analysis of the same problem where the
risk of progressive failure is conceptually addressed.
Sfriso A.O.
