Calibration of advanced soil constitutive models for liquefaction studies: application to an urban area on the eastern Sicily (Italy)

Abstract

Earthquake-induced liquefaction is considered as one of the most destructive natural phenomena for civil engineering structures causing high economic losses. In the last three decades, numerical methods, such as the Finite Element Method (FEM) and the Finite Difference Method (FDM), have increasingly been used to model the soil behaviour under cyclic loading. In this framework, the paper presents the procedure used for the calibration of advanced soil constitutive models for liquefaction studies based on in-situ investigation performed in an intensely urbanized area located on the eastern tip of Sicily (Italy). Eastern Sicily is considered as one of the Mediterranean areas most exposed to earthquakes due to the large seismic events that hit the area in the past, such as the 1169, 1693, 1793 and 1908 earthquakes. As reported by historical sources, these seismic events induced liquefaction phenomena along the coastal area. Liquefaction evidence were found in the Holocene deposit located in the eastern tip of Sicily where an intense geotechnical investigation, including in situ and laboratory tests, have been carried out in order to define the subsoil model. In this study, numerical aspects of liquefaction have been captured using two different constitutive models that are the UBC3D-PLM model and the PM4Sand model implemented in the finite element code PLAXIS and in the finite difference code FLAC, respectively. The parameters of both constitutive models were calibrated to the relationship proposed by Boulanger and Idriss (2014) for the cyclic resistance ratio based on SPT data, CRRMw=7.5, σv0΄=1atm, by the simulation of cyclic direct simple shear tests (CDSS). The results showed a satisfactory fit between the calibrated models and the Boulanger and Idriss liquefaction triggering curve.

In view of this convergence, this parametrical fitting provides a useful information for modelling profiles where liquefiable layers condition the behaviour of critical infrastructures in the area under consideration characterized by high seismic risk.