1923 Gleno Dam Break:
Case Study and Numerical Modeling
Marco Pilotti; Andrea Maranzoni; Massimo Tomirotti; and Giulia
Valerio
Università degli Studi di Brescia - Brescia (IT) e-mail:
marco.pilotti@ing.unibs.it
versione completa in formato PDF
Introduction
Despite technological advances in recent decades, the worldwide safety of dams
does not appear to have greatly improved since the major failure rate in the
period between 1946 and 1955 almost doubled in the following decade (Johnson and
Illes 1976; Singh 1996), with a rising trend toward the end of the last century
(Donnelly and Morgenroth 2005; Saxena and Sharma 2004).
At the same time, several survey and monitoring studies (e.g., Douglas et
al. 1998) have reckoned that the annual probability of a major dam failure is in
the order of 10 4, which is one order of magnitude lower than the probability
usually assumed for the design of spillways. Accordingly, in spite of its very
small probability of occurrence, dam failures are, to a certain extent,
unavoidable, as a consequence of structural deficiencies (approximately 60% of
the recorded accidents), inadequate spilling design or inaccurate design flood
estimation (30%), and, finally, landslides that can cause dam overtopping in the
reservoir (10%) (Singh 1996).
In the future, climate change and progressive structural deterioration of the
oldest dams might play a significant role in modifying these percentages, while
the continuous and rapid growth of built-up zones in the tailwater areas will
significantly increase the exposure level.
Given this state of affairs, the problem of dam safety assessment is of
considerable importance, and in many countries national technical legislation
requires delineation of areas that would be flooded in the event of dam failure.
Accordingly, predicting the flood propagation effects following a hypothetical
dam failure is an indispensable requirement both for land-planning purposes and
for the formulation of disaster contingency plans.
In this direction, the approach widely accepted is based on the shallow water
equations (SWE), since they best combine computational efficiency and accurate
reconstruction of the physical phenomenon, as clearly shown by recent scientific
literature (e.g., Soares Frazão et al. 2000, 2003) produced by the important
European projects Concerted Action on Dambreak Modelling (CADAM) and
Investigation of Extreme Flood Processes and Uncertainty (IMPACT). In
particular, general agreement exists that SWE can be used to describe dam-break
waves even when they take place over real topography (Hervouet and Petitjean
1999; Valiani et al. 2002; Begnudelli and Sanders 2007; Aureli et al. 2008b;
Natale et al. 2008; Pilotti et al. 2010). However, in the presence of very
irregular topography characterized by steep bed slopes as in mountain regions,
the key SWE assumptions of small bed slopes and negligible vertical acceleration
(e.g. Cunge et al. 1980; Toro 2001) are violated. Therefore, in this context it
seems important to evaluate their applicability (Pilotti et al. 2006, 2010). [...]