Seismic performance of slopes can be assessed through displacement-based procedures
where earthquake-induced displacements are usually computed following Newmark-type
calculations. These can be adopted to perform a parametric integration of earthquake
records to evaluate permanent displacements for different slope characteristics and seismic
input properties. Several semi-empirical relationships can be obtained for different
purposes: obtaining site-specific displacement hazard curves following a fully-probabilistic
approach, to assess the seismic risk associated with the slope; providing semi-empirical
models within a deterministic framework, where the seismic-induced permanent displacement
is compared with threshold values related to different levels of seismic performance;
calibrating the seismic coefficient to be used in pseudo-static calculations, where a safety
factor against limit conditions is computed. In this research project, semi-empirical relationships will be
obtained as a result of a parametric integration of an updated version of the Italian strong motion
database, that, in turn, is described and compared to older versions of the database
and to well-known ground motion prediction equations. Permanent displacement will be
expressed as a function of either ground motion parameters, for a given yield seismic coefficient
of the slope, or of both ground motion parameters and the seismic coefficient. The
first are meant to be used as a tool to develop site-specific displacement hazard curves,
while the last can be used to evaluate earthquake-induced slope displacements, as well as to
calibrate the seismic coefficient to be used in a pseudo-static analysis. Influence of the vertical
component of seismic motion on these semi-empirical relationships will also be assessed.
In this research project, Newmark-type computations will be performed through a parametrical integration of an updated version of the Italian seismic database. This includes seismic events recorded in the Italian territory in the time frame 1972-2017, thus including the destructive 2009 L'Aquila, 2012 Emilia and 2016 Central Italy earthquakes: this constitues the first and main aspect of novelty of the research project at hand.
Moreover, it will be seen that ground motion parameters characterising the new Italian seismic database, such as the peak ground acceleration PGA, the Arias intensity IA, the mean period Tm and the significant duration D5-95, can be reasonably computed using, on a first approximation, well-known ground motion prediction equations (GMPEs) already available in the literature. It is understood that the new database can be used to calibrate new GMPEs as well, and this is the second aspect of novelty of the study.
Furthermore, both one- and two-ground motion parameter, semi-empirical models will be developed following Rathje and Cho (2019), where the permanent displacements are expressed as a function of some parameters of the seismic motion, for a given yield seismic coefficient ky. Two-ground motion parameter relationships will be certainly characterised by a higher coefficient of determination R2 when accounting for the frequency content and the duration of the seismic motion, in addition to its intensity: among these, the best model, both in terms of accuracy and ease of use, will proved to be the (PGA, PGV) model. These semi-empirical models can be used to evaluate seismic-induced displacements and, in turn, the seismic performance of a slope characterised by a given ky, through the comparison with threshold displacements dy. The ones already available for the Italian National territory can be deemed outdated, and therefore this is a third aspect of novelty of the study.
The semi-empirical relationships can be modified to explicitly introduce the dependency of slope displacements on the ratio between the yield and the maximum seismic coefficient, ky/kmax. These relationships can be used to either estimate the permanent displacement d induced by earthquake loading or to evaluate the seismic coefficient k to be adopted when using a pseudo-static approach. To this end, upper-bound (94th percentile) relationships linking the permanent displacement d to the ratio ky/kmax will be proposed and values of the seismic coefficient computed assuming threshold displacements dy = 2, 5 and 15 cm, three subsoil groups (rock, stiff and soft soils) and four acceleration levels (PGA = 0.05, 0.15, 0.25 and 0.35 g). It is worth mentioning that the procedure laid out in this study to compute the equivalent seismic coefficient can be readily adopted for different values of threshold displacements dy and for different percentiles: this is a crucial aspect of this research project, as this provides both results to be readily used and a procedure to be followed in the future by other researchers.