Un metodo disaccoppiato per la valutazione del rischio di liquefazione e la sua mitigazione tramite drenaggi verticali di ghiaia

Proper evaluation of seismic-induced excess pore water pressures in saturated sandy soils is still an open issue. Despite semi-empirical methods being commonly employed in the engineering practice to evaluate the liquefaction hazard, based on either SPT or CPT results, a series of limitations have been identified in recent years. In particular, it was found that the undrained assumption during a seismic event is not realistic for sandy soils, as well as the influence of non-liquefiable layers on the occurrence and manifestation of liquefaction is now clear. Additionally, the importance of evaluating excess pore pressure values prior to complete liquefaction is evident, as the reduction of effective stress may provide ultimate limit states in the soil. Therefore, a more reliable approach is necessary, which may be capable of providing the excess pore pressure time histories across the entire layered soil domain. Starting from the work by Seed et al. (1975), this study presents a novel uncoupled method for computing seismic-induced excess pore water pressures developing in a 1D layered soil column under partially drained conditions. In particular, new simplified assumptions are introduced to account for physical aspects such as soil stiffness degradation, modification of the frequency content of ground motion due to pore pressure build-up, and the representation of the input motion through a more reliable procedure. Moreover, the proposed uncoupled approach was extended to the axisymmetric case in the presence of a gravel drain to investigate liquefaction mitigation, given that standard design methods of the gravel drains are typically based on the seminal work by Seed & Booker (1976), which relies on several simplifying assumptions. The uncoupled approach, both for the 1D and axisymmetric configuration, was implemented in Matlab via the Finite Difference Method and validated against extensive numerical parametric studies, which were based on fully-coupled FE analyses. The numerical analyses were crucial for a full understanding of the phenomenon, both in the configuration with and without the gravel drain, highlighting the influence of the different governing parameters and the physical aspects that should be accounted for in the simplified methods for improving their reliability.